CA2162131A1 - System and method for determining the impact of weather and other factors on managerial planning applications - Google Patents

Abstract

A computer-based Executive Information System (202) for determining the impact of weather and other external and internal factors on the retail industry. Utilizing a multiple regression correlation technique in a predictive model, a correlation of weather variables with store information for specific locations and times is performed to quantify a weather impact model in terms of unit or dollar sales volume change, or any other commercially useful benchmark (704). The Executive Information System determines these relationships with location and time specificity. Using the relationship between historical weather and historical sales, the system "deweatherizes" the historical weather to create a normalized historical sales relative to weather based on the 30 years average normal weather (720). The deweatherized data may be used in conjunction with a user-provided managerial plan to produce a revised managerial plan (207). Alternatively the revised plan can be "weatherized" by the system by applying forecasted weather to the weather impact model to generate a weather-modified managerial plan (204).

Description

System and Method For Determinin~
The ~mp~^t of Weather and Other Factors On Managerial Pl~nnin~ Applications .

Background of the Invention Field of the Invention The present invention relates generally to predicting consumer demand ;lllS relative to the retail industry, and more particularly, to identifying theimpact of weather and other factors on retail sales.

RP7~7te~ Art A. Historical P~ e~h~e of RPt~ 7g The retail industry has historically been influenced by the shape of the times. For example, the retail industry is imp~( ted by war and peace, lifestylechanges, demographic shifts, attitude progressions, econo~llic e~p~n~ion and contraction, tax policies, and ~;ul~ell~;y fl~lct..~tions.

The period from 1965 to 1975 was marked by growth and segmentation in the retail industry. New types of stores such as department stores, specialtystores, and discount stores appeared, increasing colll~eLiLion in the retail industry. One result of this growth was a decrease in gross margin (sales price - cost of goods sold). Another result was a shifting of supply sources.

Originally, merch~n~ e was supplied exclusively by vendors. However, segmentation and growth resulted in specialty chains and discounters m~nnf~rtllring merch~ntli~e in-house (commonly known as vertical integration).

S The period from 1975 to 1980 was marked by di~ lcionment and complexity in the retail industry. Inflation and women entering the work force in signific~nt numbers resulted in a more sophi~tir~trcl consumer. Many retailers began to rethink the basics of merçh~n~ ing in terms of merc.h~n~i~e assortments, store ~lt;sellL~tions, customer service, and store locations. Otherless sophi~tir~ted retailers continued on an undisciplined and unstructured policy of store growth.

The period from 1980 to 1990 was marked by recovery and opportunity in the retail industry. An econolllic boom stim~ ted consumer confidenre and clem~nrl This, coupled with the e~p~n~ion of the previous 1~ period, paved the way for the retail industry to overborrow and overbuild.
With their increased size, retailers became increasingly unable to manage and analyze the h-rollll~ion flowing into their org~ni7~tions.

B. Ret~ g Problems and Opportl~ities of Today The problems and opportunities facing the retailer fall into two categories of factors: (1) external factors; and (2) internal (or industry) factors.
External factors impacting the retail industry include, for example, adverse or favorable weather, rising labor costs, increasing property costs, increased competition, economics, increasing cost of capital, increasing consumer aw~lt;ness, increasing distribution costs, çh~nging demographics and zero population growth, decreasing labor pool, and flat to ~limini~hing per capita income.

WO 9S/24012 2 1 6 2 1 ~ 1 PCT/US95/02557 Internal (or industry) factors affecting the retail industry include, for example, large number of stores (decentrali_ation), homogeneity among retailers, continuous price promotion (equates to decreased gross margin), decreasing customer loyalty, minim~l customer service, physical growth limitations, and large qll~ntities of specific retailer store inro~ aLion.

Growth and profitability can only be achieved by m~ximi7ing the productivity and profitability of the primary assets of the retail business:
merch~n~ e (i~venLoly), people, and retail space. The above external and industry factors have added to a retailer's burdens of m~int~ining the productivity of these assets.

Of the three primary assets, merrh~n~ e productivity is particularly important due to the limiting effect of external and internal factors on people and space productivity (e.g., physical growth limitations and high labor costs).Merch~n~ e productivity can be best achieved by m~int~ining effective mix of product in a store by product cha,~cLe-istic (merch~ndi~e assortments).

To achieve more err~;~ive mer~h~ li.ce assortments, a retailer must have a merch~n.li~e plan that provides the retailer with the ability to (1) define, source, acquire, and achieve specific target merc.h~n~ e asso.L...e..l~
for each individual store location; (2) achieve an efficient, non-di~. ~live flow from supply source to store; (3) m~int~in store assortments which achieve anticipated fin~nri~l objectives; and (4) co"""~ ir~te err~Liv~ly across all areas of the business to facilitate coordinated action and reaction.

Such an erreclive merch~n~i~e plan must consider all possible external and industry factors. To obtain this knowledge, a retailer must have responsive and easy access to the data associated with these factors, referred to as external and industry data, respectively. To ~simil~t~ and analyze this data, which comes from many sources and in many formats, retailers began W09g/240l2 2 1 6 2 1 3 1 PCT/US95/02557 ~

utili7ing management information systems (MIS). The primary function of the MIS department in the retail industry has been the electronic collection, storage, retrieval, and manipulation of store information. Mainframe-based systems were primarily utilized due to the large amount of store in~r,.,~ion S generated. Store illro,ll~alion includes any recordable event, such as purchasing, receiving, allocation, distribution, customer returns, merch~n~lice transfers, merchan~ e markdowns, promotional markdowns, inventory, store traffic, and labor data. In contrast to the extensive collection and storage of internal data, these systems, did not typically process external data. Rather, this non-industry data was simply gathered and provided to the retailer for personal interpretation.

Since underst~n-ling of-local and region level dynamics is a requisite for increased retailing productivity, retailers would escenti~lly feed store illro~ Lion at the store level into massive mainframe ~latah~ces for subsequent analysis to identify basic trends. However, the use of m~infram~s typically requires the expense of a large MIS department to process data requests.
There is also an inherent delay from the time of a data request to the time of the actual e~eclltit)n. This structure ~lc;vell~ed MIS systems from becoming cost erre~iLive for use by executives in making daily decisions, who are typically not co-npu~er speci~lictc and thus rely on data requests to MIS
specialists.

Figure 1 illu~Ll~LL~s a block diagram of a conventional MIS system architecture used in the retail industry. Referring to Figure 1, an MIS
architectl-re 101 captures store information (one form of internal data) and electronically flows this information (data) throughout the organi7~tion for managerial planning and control purposes.

At point of sale 104, scanners 108 and electronic registers 110 record trancartions to create POS data 106. These tranc~ctions include data related _5 _ to customer purchases, customer returns, merçh~nrli~e transfers, merch~n~i.ce markdowns, promotional markdowns, etc. POS data 106 is one form of store inrollllation 116. Store inrollll~Lion 116 also includes other store data 112.
Other store data 112 includes data related to receiving, allocation, distribution, inventory, store traffic, labor, etc. Other store data 112 is generally gener~edby other in-store systems.

Store info.lllaLion 116 is polled (electronically transferred) from point of sale 104 by he~ u~Lel~ typically by modem or leased-line means 117.
POS 104 ~ senL~ one typical location (retail store). However, MIS
arçhitectnre 101 can support multiple POS locations 104.

A data storage and retrieval facility 120 receives store information 116 using co~ uler hardware 122 and software 124. Data storage and retrieval facility 120 stores store il-r~,llllaLion 116. Store hlfolll,aLion 116 is retrieved into data analyzer 127. Data analyzer 127 shapes and analyzes store inro.~ Lion 116 under the command of a user to produce data, in the form of reports, for use in the l~.~al~Lion of a managerial plan 130.

In the 1970's and 1980's, retrieval of store inr(j..ll~Lion 116 into data analyzer 127 and the subsequent report genel~Lion were m~ml~lly or electronically gene-,lL~d through a custom request to MIS department personnel. More recently, in response to the need for a rapid executive interface to data for m~n~gerial plan preparation, a large industry developed in Executive Inrc,....aLion Systems (EIS). Referring to Figure 1, an EIS 129, which typically ope-~es on a personal co".puLer ~o,h~LaLion platform, interfaces with the MIS m~inframe or mid-range rl~t~b~e in data storage and retrieval facility 120. An EIS system is a co.-lpuLer-based system by which inro.l.~aLion and analysis can be ~çces~ed, created, p~ ged and/or delivered for use on dem~ntl by users who are non-teçhnic~l in bacl~loulld. Also, EIS
systems perform specific managerial applications without extensive interaction with the user, which reduces or elimin~t,os the need for co~..puL~r software training and documentation.

In contrast to store inro----ation 116, external inrol,lla~ion 136 consists of manual reports covering such topics as economic forecasts, demographic changes, and competitive analysis. In conventional systems, external hlrol"~Lion 116 iS separately made available to the user for consideration in developing managerial plan 130.

Technic~l improvements in speed and storage capability of personal computers (PCs) have allowed this trend towards EIS systems to take place, while most firms still m~int~in a m~infr~me or minicon.~u~t;r architect--re for basic POS data storage and proces~ing. The advent of powerful mini coll,pu~ , local area ne~works (LANs), and PC systems has resulted in many of the traditional mainframe retailing applications migrating to these new platforms.

C. The Nature of Weather~Poma~ies Weather anomalies are more of a regional and local event rather than a national phenomenon in countries as geographically large as the United States. This is not to say that very anomalous weather cannot affect an entire country or continent, creating, for example, abnormally hot or cold seasons.
However, these events are less frequent than regional or local aberrations.
Signifi~nt precipitation and It;--,pe,~ture deviations from normal events occur continually at time intervals in specific regions and locations throughout the United States.

Rec?~ e actual daily oc-;u~llces fluctuate around the long term "normal" or "average" trend line (in meteorology, normal is typically based on a 30 year average), past historical averages can be a very poor predictor WO9S/24012 2 1 6 2 1 3 1 PCT/US9~;/02557 of future weather on a given day and time at any specific location. Implicitly, weather effects are already embedded in an MIS POS l~t~h~e, so the retailer is consciously or unconsciously using some type of historical weather as a factor in any planning approach that uses trendline forecasts based on historical POS data for a given location and time period.

D. Weather Relative to National P~n~1ni-lg Applications At a national level, weather is only one of several important variables driving consumer demand for a retailer's products. Several other factors are, for example, price, colllpe~ilion, quality, advertising exposure, and the structure of the retailer's operations (number of stores, square footage, locations, etc). Relative to the national and regional implement~tiQn of planning, the impact of all of these variables domin~t~s trendline projections.

As described above, POS ~t~h~es track sales trends of specific categories at specific locations which are then aggregated and manipulated into regional and national executive i~lrol~l~aLion reports. Since the impact of local weather anomalies can be diluted when aggregated to the national levels (sharp local sales fluctuations due to weather tend to average out when aggregated into national numbers), the impact of weather has not received much scrutiny relative to national planning and r~Jlcç~l;ng.

E. WeatherRelatiYe to Regional and Local P~rrni~gApplications The impact of weather on a regional and local level is direct and dramatic. At the store level, weather is often a key driver of sales of specificproduct categories. Weather also influences store traffic which, in turn, often impacts sales of all goods. Weather can influence the timing and intensity of markdowns, and can create stockout situations which replenishment cycles can not address due to the inherent time lag of many repleni~hmPnt approaches.

WO 95124012 2 1 6 2 l 3 1 PCT/US9~/025S7 The combination of lost sales due to stockouts and markdowns required to move slow inventory are enormous hidden costs, both in terms of lost income and opportunity costs. Aggregate these costs on a national level, and weather is one of the last major areas of retailing where costs can be carved S out (elimin~te overstocks) and stores can improve productivity (less markdown allows for more margin within the same square footage).

In short, weather can create windows of opportunity or potential pitfalls that are completely independent events relative to economics, demographics, consumer income, and competitive issues (price, quality). The cash and opportunity costs in the aggregate are enormous.

F. Conventional Approaches Addressing Weat*er Impact Though the majority of retailers acknowledge the effects of weather, many do not consider weather as a problem per se, considering it as a completely unco~ ollable part of the external environment.

However, the underlying problem is essenti~lly one of prediction of the future; i.e., developing a predictive model. All retailers must forecast (informally or formally) how much inventory to buy and distribute based on expected demand and ~l l~,iaLe illvelllOly buffers. Hence, many conventional predictive modelling processes have been developed, none of which adequately address the impact of weather impact.

One collvenLional solution is to purposely not consider the impact of weather on retail sales. In such inct~n~es, the retailer will m~int~in high inventory levels and rapidly replenish the illvellLoly as it is sold. This approach creates large working capital needs to support such a large illv~lLo~y.

_9_ Another conventional solution is for the retailer to qualitatively use weather information to anticipate future demands. This procedure, if used by decision makers, is very subjective and does not evaluate weather in a predictive sense. Nor does it quantify the effect of past and future weather on consumer demands.

Another conventional approach is the utilization of climatology.
Climatology is the study of the climates found on the earth. Climatology synth~si7es weather elements (temperature, precipitation, wind, etc.) over a long period of time (years), res.llting in characteristic weather patterns for agiven area for a given time frame (weekly, monthly, seasonably, etc.). This approach does not utilize forecasted weather as a parameter, which can vary considerably from any given time period from year to year for a given area.
Climatology yields only the average weather condition, and is not indicative of the weather for any specific future time frame.

~ l r~ , e~ ~, and retailers have been known to rely on broad projections developed by the National Weather Service (the govcllllllental entity in the USA charged with dissemin~ting weather data to the public) and other private forec~ting firms. With ,cfelcllce to long range projections, these may be vague, broad, and lack regional or local specificity. It is of limited use since they are issued to cover ~nticip~t~ weather averaged for 30, 60, or 90 day periods covering large geographic areas. This inl~ ion cannot be qll~ntifie~l or easily integr~ttqd into an MlS-based planning system which is geared toward a daily or weekly time increment for specific location and time.

In sl-mm~ry, the above collven~ional solutions to weather planning problems in retail all suffer from one or several deficiencies which severely limit their commercial value, by not providing: (1) regional and/or local specificity in m~ nring past weather impact and projecting future weather WO9S/24012 2 1 6 2 1 3 1 PCT/US95/025!i7 ~

impact, (2) the daily, weekly, and monthly increment of planning and forer~.cting required in the retail industry, (3) ample forecast leadtime required by such planning applications as buying, advertising, promotion, distribution, financial budgeting, labor schto~ ng~ and store traffic analysis, (4) the qll~ntific~tion of weather impact required for precise planning applications such as unit buying and unit distribution, financial budget forec~cting~ and labor sch~ lling, (5) reliability beyond a 3 to 5 day le~ltime, (6) a predictiveweather impact model, which links quantitative weather impact measurement through historical correlation, with qu~ntit~tive forecasts, (7) the ability to remove historical weather effects from past retail sales for use as a baseline in sales forec~.cting, (8) an entirely electronic, col-lpul~riz~d, EIS
implementation for ease of data retrieval/analysis with specific functions that solve specific managerial planning applications, and (9) a graphical user interface l~c;sel~Ling a predictive model in graphs, formats, and charts immefli~tely useful to the specific managerial applications.

What is needed is an Fxecutive Inrc.. aLion System (EIS) which contains a predictive model lltili7:ing weather and other external and internal factors to provide location and time specific forecasts. The forecast must be available early enough to provide the nt~.cesc~ry lead time for retail planners to respond to the ch~nging factors, and must be reliable. The EIS system must interface to exicting MIS systems, and must present the analysis quickly and in a form which is tailored to specific planning applications.

WO 9~/24012 2 1 6 2 1 3 1 PCT/US951025~i7 Summary of the Invention The present invention is a computer-based EIS system which determines the impact of weather and other external and internal factors on the retail industry. The present invention is a long-range executive weather inro.l,-alion system (LEWIS) cont~ining a predictive model which utilizes industry and non-industry data in its determination of the relationship between historical weather and retail sales. LEWIS determines these relationships with location specificity (for example, store level) as well as time specificity (forexample, daily and weekly time increments).

The present invention utilizes a multiple regression correlation technique to generate a weather impact model which correlates weather and other variables with store infcllllation for specific locations. The weather impact model quantifies the weather impact in terms of unit or dollar sales volume or any other commercially useful benchm~rk After dele,lllhling the relationship between historical wealher and historical sales, LEWIS ~,enPli.lrs a norm~li7ed or deweatherized historical sales baseline utili7ing normal weather and external factors. Normal weather is defined as the 30 year average of a particular weather ~ er for a specific location and time.

Since the original managerial plan does not typically consider weather effects in any ~y~Le.-.atic function, a user can utilize the deweatherized data to ge.lelaLt; a revised managerial plan. In addition, LEWIS may then be implemt~ntlq-i to "weatherize" the revised managerial plan. That is, LEWIS
utilizes the previously determined relationships and applies them to a forecasted weather to gene,d~e a weather-modified managerial plan.

WO 95/24012 PCT/US95/02~57 A weather impact model for buying, distribution, financial budgeting, labor schr~ ling, advertising, promotion, and store traffic analysis applications, is used in conjunction with 1) daily and weekly weather forecasts, and 2) a managerial plan forecast for a specific time, product, and location. The resulting output represents a weather-modified plan for the above managerial planning applications. This weather modified managerial plan is manipulated by a graphic user interface (GUI) into useful charts, graphics and reports for rapid ~cimil~tit~n by the user.

For advertising and promotional applications, the weather impact model is used in conjunction with store information and forecasted weather data. The output identifies how products are favorably or unfavorably imp~rte(l by weather at any given location (that is, cold impact versus hot), the degree of the impact, and most importantly, the most favorable timing for advertising and promotional c~mp~ign~ to take advantage of weather impact.

The present invention provides (1) regional and local specificity in measuring the impact of past weather and projecting the impact of future weather, external, and internal factors, (2) daily, weekly, and monthly increment of planning and forec~ting, (3) sufficient forecast l~(ltime required by such planning applications as buying, advertising, promotion, distribution, fin~nri~l budgeting, labor schedllling and store traffic analysis, (4) the q--~ntifir~tion of weather impact required for precise planning applications such as unit buying and unit distribution, financial budget forec~ting, and labor sçhed--ling, store traffic, advertising, and promotion, (5) reliability beyond a 3 to 5 day Iç~-ltime, (6) a predictive weather impact model, which links quantitative weather impact measurement through historical correlation, with qu~"~ l;ve forecasts, (7) the ability to remove historical weather effects from past retail sales for use as a baseline in sales forec~ting, (8) an entirely electronic, con,~uL~lized, EIS implementation for ease of data retrieval/analysis with specific functions that solve specific managerial .

planning applications, and (9) a graphical user interface represents the predictive model in graphs, fo,ll~aL~, and charts imme~ tely useful to the specific managerial applications.

Further fealures and advantages of the present invention, as well as the structure and operation of various emb~-liment~ of the present invention, are described in detail below with reference to the accolllpanying drawings. In the drawings, like rerer~llce numbers in-lir~te identical or functionally similar elements. Additionally, the left-most digit of a lerelellce number identifies the drawing in which the lt;Çe,c;llce number first a~pea~.

WO 9S/24012 2 1 6 2 1 3 1 PCT/US9~/02557 Brief Description of The Drawings This invention will be better understood if rere,t;llce is made to the accompanying drawings in which:
Figure 1 is a block liAgrAm of a typical management information system (MIS) archit~-ct~lre;
Figure 2 is a interface block (li~grAm showing the Long-range Executive Weather Inro""~Lion System (LEWIS) creating and modifying managerial plans;
Figure 3 is a timeline illustrating the Iç~lltime required to execute specific elements of a managerial plan;
Figure 4 iS a block diagram showing an MIS system cont~ininp: the LEWIS system of the present invention;
Figure S illustrates some types of store and external inrc",l~alion;
Figure 6 is a block diagram a local area neL~olk cont~ining ~olh~L~Lions ~e~Ço""ing mAnAgeriAl planning functions lltili7:ing LEWIS;
Figure 7 is a block diagram illll~tr~ting the data flow and analytical process of LEWIS;
Figure 8 is a flowchart of the functions pelro",led by the ~t~bA~e initi~ er;
Figure 9 is a block rli~gr~m of the functions pe,ro""ed by the correlation processor; and Figure 10 is a flowchalL of the functions ~elrol"léd by the forec~ting processor.

WO 95/24012 2 1 6 2 1 3 I PCT/US95/02~57 .

Defni7e~ Description of the Preferred Embodiment~

I. General Overview The present invention is an Executive Information System (EIS) for managerial planning applications that are imp~rt~d by weather, economics, competition, demographics, and other factors. The present invention, referred to as a Long-range F.xeclltive Weather Infoll-lation System (LEWIS), calculates and displays the impact of any internal (industry) and external factors on retail sales pe-r~,-...ance. This calculation includes the modeling of the relationship be~ween the above factors and retail inro.. ation. LEWIS
utilizes such a model in conjunction with long range weather forecasts to identify the future impact of the above factors on retail planning applications.
In the plefellt;d embodiment of the present invention, LEWIS is implemented to deLell,lille the impact of various factors on the retail sales industry. However, as one of ordinary skill in the relevant art would find a~lJa,cll~ the present invention may be implemented to apply to any industry affected by weather. For example, the present invention may be implemented in industries such as the personal care, utilities~ energy, agriculture, cons,llller products, in~llr~nre, ~ ol ~ion, outdoor events, chemicals, pharmaceutical, construction, entert~inm.o-nt and travel industries.

Figure 2 illustrates a block diagram of an implementation of the present invention in a retailing envi,ol----ent. Referring now to Figure 2, LEWIS 202 receives as inputs external inr(jlllla~ion 136, store inroll-~a~ion 116, and weather data 201. LEWIS 202 utilizes these inputs to produce deweatherized data 205. Deweatherized data 205 may be utilized by a user as a new b~elin~-in developing managerial plan 130. In a p~rellc;d embodiment of the present invention, managerial plan 130 is any weather-il.lp~cled managerial plan or WO 95/24012 PCT/US9~/025~7 21 62 1 31 ~

analysis which can accept the level of deweatherized data 205 produced by LEWIS 202. Examples of managerial plan 130 include buying product 206, distributing product to specific locations 208, advertising 210, promotions 212,~ln~nri~l budgeting 214, historical analysis 220, labor sçhPdnling 216, and store traffic analysis 218. These applications are by way of example only; the present invention is intended to apply to any managerial application that experiences weather impact in any industry. In a preferred embodiment, a user modifies one or more portions of managerial plan 130 in light of deweatherized data 205 as part of an original planning or replanning function.
The subsequent managerial plan is referred to as a revised managerial plan 207.

If a user cannot create a managerial plan which will be able to accept the level of deweatherized data gen~r~tecl by LEWIS 202 ("level" mP~ning by product, by store, and by week), deweatherized data 205 is retained in LEWIS
202 and utilized to internally gene~al~ a weather-modified m~n~gerial plan 204. For example, some retailers plan at the monthly or seasonal level, rather than at the weekly level. Such a retailer could not directly use deweatherized data 205.

The term "deweatherize" refers to the process by which LEWIS 202 utilizes a multiple regression technique ((ii~cu~ed below) to re-state retail data based upon normal weather and external factors. Normal weather is defined as the 30 year average of a particular weather p~r~meter for a specific locationand time. External factors imp~ting the retail industry include, for example, rising labor costs, increasing property costs, increased col..peLiLion, econo---ics, increasing cost of capital, increasing consumer awareness, increasing distribution costs, ch~nging demographics and zero population growth, decreasing labor pool, and flat to ~limini~hing per capita income. Internal (or industry) factors affecting the retail industry include, for example, large number of stores (decentralization), homogeneity amongst retailers, continuous WO 9S/24012 2 1 6 2 1 3 1 PCT/US9~i/02557 .

price promotion (equates to decreased gross margin dollars), decreasing customer loyalty, minim~l customer service, physical growth limitations, and large qu~ntiti~s of specific retailer store inro~l"aLion.

Figure 3 illustrates the typical le~-ltimes required to execute specific elements of a managerial plan in a large retailer environment. Inherent in the concep~ of managerial planning is a le~dtime associated with specific activitieswhich must be conducted in an org~ni7~d fashion.

The m~ximnm and minimllm Iç~-ltimes for the functions depicted in managerial plan 130 are illl-ctr~ted according to the time they have to be pelrolllled relative to the implement~tion date of the managerial plan. The duration of the le~dtimes illustrated in Figure 3 are those in which the planning activity is to occur to have commercial value.

Buying 206 has a typical le~-ltim~ 302 of 6 to 15 months. Distributing 208 has a typical le~-ltime 304 of 1 week to 3 months. Fin~nr.i~l budgeting 214 has a typical le-~(itime 306 of 3 to 12 months. Advertising 210 and promotion 212 applications are con-luctçd at different levels of retailing such as regional and national levels. As a result, the Iç~-ltime required to implement these applications varies according to the targeted level. For example, regional advertising requires a lç701tim~ 308 of 1 to 3 months, while national advertising requires a le~dtime 310 of 3 to 12 months. Regional promotion requires a le~-ltime 312 of 3 days to 3 months, while national promotions require a le~dtime 314 of 1 to 6 months. Store traffic analysis 218 has a Iç~-ltime 316 of 3 days to 12 months, and labor schsdllling 216 has a le~dtime 318 from 3 days to 3 months. Historical Analysis 220 is pelrolllled on an on-going basis and thus is shown to have a continual lç~-ltime 320.

WO 95/24012 2 1 6 2 1 3 1 PCT/US9~/02557 II. MIS Architecture Implçnlen~frlg The Present Invention Figure 4 illustrates a block diagram of a typical MIS system in which the present invention is implemented. Referring to Figure 4, LEWIS 202 is resident within a co~llpuLer-based MIS architectl~re 402. MIS architectllre 402 captures store infor,ll~Lion 116 and external hlro-.-.aLion 136 to electronically transmit this inro...laLion throughout an org~ni7~ti~n for managerial planning and control purposes in a manner similar to that described with .~rt;rel1ce to Figure 1.

Retailers sell product to consumers through one or more store locations 104 gene.~Ling store inro----aLion 116. Also note that the present invention is applicable to one or more locations, metropolitan St~ti!~tir~ll areas, or regions.
At point of sale 104, electronic scanners 108, registers 110, and other electronic sc~nning and data gathering devices gene.~Lte POS data 106. Other store data 112, which is also considered store inrorlllaLion 116, includes any type of r~co.. lable event taking place in support of consumer sales (that is, inventory transfer from distribution center to store, promotion data, store traffic, labor data, etc). Other store data 112 specifically does not include events conl;1i"il~g inro....aLion regarding the time, amount, and merch~n~ e of a specific sale. Thus, any type of store inro~ aLion 116 in support of consumer sales activities is collected and flows through MIS architect--re 402 to LEWIS 202.

Also available to the retailer is external inro~...aLion 136. External inror...ation 136 can be economic, demographic, cor-peLiLive, or any other inro~ ation that the retailer believes is of value to ~csessing his business pe-ro.. ~.ance. External information 136 is typically available via on-line dataservices or from external l~:lt~b~e sources. The data storage and retrieval facility 120 receives external information 136 and store info~ aLion 116 using co...p~Ler hardware 122 and software 124. Depending on the MIS age, scope WO 95~24012 2 1 6 2 1 3 1 PCT/US9!i/025!;7 .

and type of data, and retailer resources, the data storage and retrieval facility 120 can be a mainframe computer, mid-range coll~pul~r or Personal ColllpuLt;r (PC) network configuration. For the largest retailers, mainframe storage is preferable. However, as one of ordinary skill the relevant art would find apparent, the present invention may be implemented in any computer platform or architecture presently available or developed in the future.

In the preferred embodiment of the present invention, LEWIS 202 retrieves the appl~pliaL~ inrolllldLion from data storage and retrieval facility120 and receives weather data 201 to geneldLt; deweatherized data 205 which, used in conjunction with managerial plan 130, produces a revised managerial plan 207. In addition, revised managerial plan 207 can be input into LEWIS
202 to be "weatherized." The term "weatherize" refers to the process by which one utilizes the present invention with forecasted weather and other factors to forecast retail sales. The forecasted weather is commercially available from Strategic Weather Services, Wayne, Pellllsylv~nia.

Hence, the wolk~LdLion or LAN-based LEWIS 202 implemented within the retailer's MIS architectl.re 102 receives store inr(jlllldLion 116, externalinrollllaLion 136, and weather data 201 as inputs for analysis. The result of the weatherization yields a weather-modified managerial plan 204.

In the pl~fellt;d embodiment of present invention, the LEWIS system 202 is a colll~uLer-based F.~ec~ltive Inrolll-dLion System (EIS) residing on a PC
work~LdLion or LAN having, for example, a model 80486 processor (Intel, Sunnyvale, California, USA). However, as one of ordinary skill in the art would find apparent, LEWIS 202 may reside within any colllpuLer-based system, inchl-ling mid-range or m~infr~me MIS architectllres.

Figure 5 illustrates, in block ~ gr~m form, the store information 116 and external inrolll-aLion 136 which are considered in the prerelled WO 95/24012 2 1 6 2 1 3 1 PCTrUS95/02557 embodiment of LEWIS 202. The various categories of store inrolllla~ion 116 include POS data 106 and other store data 112. POS data 106 includes product data 502, POS receipt data 504, promotional data 510, markdown data 512. Other store data 112 includes inventory data 506, store traffic data 508 and employee labor data 516. The various categories of external information 136 include econol.lic data 503, demographic data 505, and competitive data 507.

These categories of store inrollll~Lion 116 and external information 136 are by way of example only, and the present invention colllelll~lates any ~ype of store, external, or other data collected in the course of an enterprise whichexperiences weather impact and can be analyzed to discern commercially valuable analysis for planning purposes. In industries other than retailing, thepresent invention conLGlllplates any data collected in the course of operating an enLGl~lise which is nPcç~c,.ry for and valuable to the activity of planning, including but not limited to Uniform Product Code (UPC) data, shipments to distributors, dealer ch~nn~o.l data, financial market data, labor sche~ ng data and store traffic data.

Referring to Figure 6, the ~lGrGr,ed embodiment of LEWIS 202 residing on a LAN is illu~tr~te~. In the configuration shown in Figure 6, LEWIS 202 resides on LAN 600 whereby all applications have access to LEWIS 202. In the more advanced MIS systems, data analyær 406 allows electronic transfer of managerial plans beLween planning applications residing on the same or dirrelGlll workstation. For example, wo~h~.Lation 602 may be used to perform the buying application 206 portion of the managerial plan 130.
Likewise, wo~k~.L~Lion 604 is used to perform the distribution 208 portion of managerial plan 130. Workstation 606 iS used to perform the advertising 210 portion. Workstation 608 iS used to promotion 212 portion. Workstation 610 is used to perro",l the fin~nri~l budgeting 214 portion. Workstation 612 is used to perform the labor scheclllling 216 portion, and wolL~.LaLion 614 is used WO95/24012 2 1 6 2 1 3 1 PCT~US95/02557 to perform store traffic analysis portion 218 of managerial plan 130.
Workstation 616 iS used to perform the historical analysis 220 portion of managerial plan 130.

Referring to Figure 7, a block diagram of the data flow and analytical processes resident within the LEWIS system 202 is ill-lctr~te~. In Figure 7, LEWIS 202 includes a ~t~h~ce initializer 702, a correlation ~rocessor 704, a forec~cting prucessor 706, and a graphical user interface (GUI) 710, described in detail below.

As described above, LEWIS system 202 interfaces with the retailer's MIS data storage and retrieval system 120 to receive the store and external information ill-~ctr~tP,d in Figure 5. Depending on the structure of the data fields residing in data storage and retrieval facility 120, LEWIS 202 may require the data to be aggregated or manipulated. For example, aggregating daily sales history into weekly figures which would typically be done by the retailer. This function is typically pelr~lllled prior to being input into LEWIS202.

Weather data 201 includes historical weather data 714, forecasted weather data 715 and normal weather data 716. Weather data 701 is typically measured by any time increment "eces~i1.y, for example: day, week, month.
Weather data 201 can be any p~r~mptp~r of tell-~el~tul~;, prt;ci~iLalion, hllmi~lity and other common meteorological factor.

Forecasted weather data 715 is defined as predicted weather in time in.;lt;lllenl~ for specific locations from 3 days to 15 months in the future.
Forecasted weather data 715 is commercially available from Strategic Weather Services, Wayne, Pennsylvania, U.S.A. Historical weather data 714 is defined as actual weather obsel v~Lions in time increments for specific WO 9S/24012 2 1 6 2 1 3 1 PCT~US95/02557 locations. Normal weather 716 iS defined as the 30 year average of any weather parameter.

The tl~t~h~e initializer 702 prepares store information 116 and external infollllaLion 136 received from data storage and retrieval facility 120, and S places this data into rl~t~h~es using mathematical relationships (~ c-~secl below). The ~l~t~h~e initializer 702 transforms the datasets into the proper form for the correlation processor 704. The ~1~t~h~e initi~li7~r 702 will accommodate client-specific hierarchies of products and locations and will also perru,.,~ location-to-MSA mapping functions. Data storage and retrieval facility 120 stores the store information in an ap~vl~pliate format for use by LEWIS 202 without aggregation or manipulation.

In the prere..ed embodiment of the present invention, historical store inro~ aLion 116 and external store inrollllaLion 136 are correlated with historical weather data 714. The correlation processor 704 produces the deweatheri ed data 205 based upon substit--ting normal weather data 716 into the weather impact model 720. The deweatherized data 205 iS used as a baseline input to the managerial plan 130, thereby providing the retailer with the capability to produce a revised managerial plan 207 to be used as input to the forec~ting pr~cessol 706. Alternatively, the retailer can choose to use his exi~ting managerial plan 130 as input to the foreç~tin~ process source 706 without ~ltili7ing deweatherized data 205.

The forec~tin~ processor 706 applies the forecasted weather data 715 to the weather impact model 720 in conjunction with the revised plan 207 or in conjunction with the deweatherized data 205, whichever is applupliate, based upon the level of "sophi~tic~tion" for each retailer. The term sophi~tiç~tion relates to the ability of the retailer,to produce daily and/or weekly product plans by store location. The foreç~cting processor 706 then produces a weather-modified managerial plan 204 based upon substitution of WO95~24012 2 1 6 2 1 3 1 PCT~US9~102557 forecasted weather information 715 and external information 136 into the weather impact model 720. The foreç~cting processor 706 qll~ntit~tively modifies a forecast from a managerial plan 130, or a revised plan 207, or the deweatherized data 205 and genel~tes relative co,llpalisons of weather impact on specific products at specific locations and times.

GUI 710 then receives the weather-modified managerial plan 204 from the foreç~tin~ processor 706.

In D~fo~(~re In~ z.or Figure 8 iS a flow chart illn~tr~ting the proces~ing steps which are performed by the (l~t~h~e initi~li7~.r 702 in a prc;rellt;d embodiment of the present invention.

First, in step 802, .~ h~ce initi~li7~,r 702 retrieves external inrollllalion 136 and store infolllla~ion 116 (generally referred to as client data), from data storage and retrieval facility 120. Weather data 201 is also input into l~t~h~einiti~li7er 702. Weather data 201 inclu(les historical weather data 714, forecasted weather data 715, and normal weather data 716. In the prel~lled embodiment, weather data 201 iS made available when LEWIS 202 iS in~t~
on LAN 600. The data retrieved by ~t~h~e initializer 702 iS determined by the functions and time periods selected by the user via GUI 710.

In step 804, rl~t~b~e initi~li7er 702 maps store loca~ions to metropolitan st~ti~tit~l areas (MSAs). This mapping function enables ~l~t~h~e initi~li7er 702 to determine what portion of weather data 201 iS required based upon the store information 116 provided. By mapping store location zip codes to MSAs"l~t~h~e initi~li7er 702 then utilizes only that portion of weather data 201 which is associated with those MSAs in which stores are located.

W O 95/24012 2 1 6 2 1 3 1 PCT~US9~/02557 Next, in step 806, ~l~t~h~ce initi~li7er 702 facilitates the identification/building of retailer hierarchial structure tables. Hierarchial structural tables in-lic~tr the parent-child relationship between retailer itemsand the respective levels above them. For example, bras are a child of women's lingerie which is a child of women's division which is a child of apparel, etc. The ~t~h-~ce initi~li7er 702 can consolidate the items in any manner desired by the user depending on the form of the hierarchial structural tables. Alternative consolidations wherein the user aggrega~es the information in a particular combination of products, levels, and store locations are also possible.

In step 808, ~1~t~h~e initi~li7~r 702 genel~tt;s a deweatherization regression structure file. The deweatherization regression structure file defines how LEWIS will build the weather impact model via multiple regression techniques. The deweatherization regression structure file is comprised of four sections: (1) regress columns which infliç~te how LEWIS will sample specific historical periods; (2) regression variables which in-lir~t~ how the variable ~l~t~h~e historical value is mapped into weather impact 720; (3) normal variables which in-liç~te how the variable data base normal values map into weather impact model 720; (4) variable mappings which intlir~tes what the output of the weather impact model will be named. Each of these are cu~e~l below.

The first section of the deweatherization r~glession structure file, regress columns, is the sampling of specific historical periods. Building weather impact model 720 requires that associations be made between similar historical observations of weather and sales and other external data. In order to associate the dates from store i.lro~ Lion 116 and weather data 201, store inrorl''a~ion 116 is offset into weather data file by a certain number of periods.
Database initi~li7er 702 receives this number of periods from the user to move the stored infc,l",ation 116 to achieve this ~lignment WO95t24012 2 ~ 6 ~ 1 3 ~ PCT/US9~i/025S7 .

Also, the specific periods to be sampled from both, historical weather data 714 and the historical sales data (store information) 116 are also receivedfrom the user. In the preferred embodiment, there is typically at least two year's quantity of data required.

5The second section of the deweatherization regression structure file, regress variables, contains the mappings of historical values from variable ~b~es 718 into weather impact model 720. To perru"l, this mapping function, the historical values to be used by weather impact model 720 are identified, including the transrur~ ions of those variables. These variables 10are present in sales hlroll"alion 116 and/or the weather data 201.

The following "x" vari~bles come from weather data 201, (history, normal, and forecast), and sales infor"~aLion 116 stored in data storage and retrieval facility 120.

In the pl~;fell~;d embodiment of the present invention, weather impact 15model 720 contains nine independent variables (discussed below). The first independent variable, xTEMP, is shown below.

xTEMP=TEMP-LAG(TEMP,1) This independent variable shows how LEWIS derives the first variable as a dirre,c;i~ce from the current value, TEMP, and the previous 20value,LAG(TEMP,1). This definition continues through the rem~ining independent variables, which are in(liç~tPd by the "x" in the front of them.
The function "LAG(XX,n)" refers to the lagging of the data XX by n periods.
The function "HAVERAGE(XX)" refers to deriving the shape of an historical average of the XX variable. The function "TAVERAGE(XX)" refers to 25deriving an average of the variable XX.

xTEMP(-1) =LAG(TEMP-LAG(TEMP, 1),1) xTEMP(-2) =LAG(TEMP-LAG(TEMP, 1),2) xPREC =PREC-LAG(PREC, 1) xPREC(-1) =LAG(PREC-LAG(PREC, 1),1) xPREC(-2) =LAG(PREC-LAG(PREC, 1),2) xPROMO =PRO.UNIT
xPROMO(-2) =LAG(PRO.UNIT,2) xTOT. UNIT = HAVERAGE(TOT. UNIT) Y =TOT. UNIT/TAVERAGE(TOT. UNIT) Y is the dependent variable, retail sales, which LEWIS is modeling.
The definition of this variable describes the "shape" of sales instead of the actual sales value. This shape of sales is derived by taking the actual sales value and dividing by the average sales for the entire season. For example, if the season was 3 periods long, and the actual sales values were 2, 4, and 6, the average for the season would be 4. The shape of sales would be 2/4, 414, 6/4, or .5,1,1.5.

The next section of the deweatherization regression structure file, normal variables, substitutes normal weather values in place of actual weather values that were used in the regress variables section. The following is a list of the substituted normal weather values:

xTEMP=TEMP.SEA-LAG(TEMP.SEA, 1) xTEMP(-1) =LAG(TEMP.SEA-LAG(TEMP.SEA, 1),1) xTEMP(-2) =LAG(TEMP.SEA-LAG(TEMP.SEA, 1),2) xPREC =PREC.SEA-LAG(PREC.SEA, 1) xPREC(-1) =LAG(PREC.SEA-LAG(PREC.SEA, 1),1) xPREC(-2) =LAG(PREC.SEA-LAG(PREC.SEA, 1),2) xPROMO =PRO.UNIT

WO95124012 2 1 6 2 1 3 1 PCT/US9~i/025S7 xPROMO(-2) = LAG(PRO . UNIT,2) xTOT. UNIT = HAVERAGE(TOT. UNIT) zDE-WEATH=TAVERAGE(TOT.UNIT)*RY

The line "zDE-WEATH=TAVERAGE(TOT.UNIT)*RY" shows that the output of substit~-ting normal weather into the equation generates a new shape of sales (RY from above), i.e., a deweatherized shape of sales. This shape is subsequently scaled back into units or dollars by multiplying it by theseasonal average which is derived as TAVERAGE(TOT.UNIT).

The next section of the deweatherization regression structure file, variable mappings, labels the deweatherization output variables. This is shown as:

variable ~ hlgS
r, LY.SLS=TOT.UNIT
r, TOT.UNIT=RY[TOT.UNIT]

The code "r" tells the output procedure that the codes following are only to pertain to the historical sections of the file. The code "LY.SLS=TOT.UNIT"
refers to last year's actual sales. the code "TOT.UNIT=RY[TOT.UNIT]"
refers to the deweatherized data (last year's sales deweatherized).

These outputs will be placed into a comma-sep~r~t~(l file in the form of product, location, variable, timel.. timeN, i.e., time is the across subscript dimension(same as the input files). This output file will be the deweatherized data 205, mP~ning that it will show the last year actual sales results and the deweatherized data results for each (product x location) colllbill~lion that wasrun through the weather impact model 720.

WO 9~/24012 PCT/US95/02557 IV. Correlation Processor Figure 9 is a flowchart illu~tr~ting the steps performed by correlation processor 704. Referring to Figure 9, in step 902 correlation processor 704 gene,~L~s the weather impact model based upon the deweatherization regression structure file defined in step 808. The weather impact model utilizes a multiple regression technique which is well known to one of ordinary skill in the art.

The weather impact model 720 is a multiple regression model with "k"
variables. This model is based on the assumption that there is a correlation (i.e., a St~ti~tic~lly signi~lC~nt relationship) between the change in weather (i.e., lelllpe.~ule and precipitation), and the change in sales. Multiple regression is the st~ti~tic~l technique employed by the correlation ~lucessor 704 to quantify these relationships, and to turn them into a usable equation, referred to as the weather impact model. The d~weall,e,ization regression lS model also considers other variables which are not strictly weather-based to more accurately define the observed changes in retail sales. These include the external and internal factors ~ cusced above.

The general form of the dewe~lt;lization regression model which provides the "best fit" to the observed retail sales data values is shown below.

20Y ~1 l32X2 1~3X3 ~ LXL

wherein, Y = dependent variable; change in sales X2.. Xk = independent variables; ch~nges in weather, external and internal factors Bl.. Bk = regression coefficients WO95/24012 2 1 6 2 1 3 1 PCT/US95/02~i~i7 .

Weather impact model 720 is essentially this equation with the values of the coefficients determined, since these coefficients identify the effect of weather on the dependent variables (retail sales).

There are a variety of techniques which can be employed to determine the regression coefficients. These techniques are considered to be obvious to one of ordinary skill in the relevant art. A more detailed description of the st~ti~tiç~l methods employed to determine regression coefficients may be found in "Econometric Models and Economic Forecasts," authored by R.S. Pindyck and D.I. Rubinfeld, the relevant portions of which are herein incorporated by reference.

The determination of variable transr~,lllla~ions, which determines how variables are to be used within the It;gles~ion equation, is critical to the sl1cce~rul execution and use of the r~lt;ssion equation. Therefore, the variable transrollll~Lions in an equation structure are shown above. These tran~rollll~ions are genel~ed by ~t~k~e initi~li7er 702.

The regression model has k + 1 variables - a dependent variable, and k independent variables (which includes a constant, shown as B1 in the equation above). There are also N obselv~ions. We can ~..,.,...~. ize the ~,lt;ssion model by writing a series of equations, as follows:

Yl = ~1+~2X21+~3X31+~4X41+--- +~X~+~l Y2 = ~ 1 + ~2X23 + ~3X32 + ~4X42 + -- + ~X~ + ~2 ... ... + ... + ... + ... + ... + ... + ...
YN ~1 ~2X2N ~3X3N ~4X4N -- ~ tN ~N

The collt;~onding matrix formulation of the model is:

Y=X~ +~

in which Yl 1 X21 - X~ -1-Y 1 X22 -- X~ ~ = ~2 =
... ... ... ... ... ... ...
YN 1 X2N - X~N. -~ ~N

where Y = Nxl column vector of dependent variable observations S X = Nxk matrix of independent variable observations ,B = kx1 column vector of unknown parameters Nx1 column vector of errors The technique for solving the deweatherization lc~ s~ion equation using matrix manipulation is:

~ (X X) (X

Given the two m~trices, X of order m x n and Y of order n x p:

Xll X12 Xl~ Yll Yl2 -- Ylp X = X2l X22 -- X2~ and y = Y2l Y22 -- Y2p ... ... ... ... ... ... ... ...
- Xml Xm2 ~ X~a ~ Ynl Yn2 ' Ynp the result of the multiplication is:

XllYll X12Y21 ''' XlaYal XllY12 X12Y22 ''' X2aYa2 ~ XllYlp X12Y2P ''' XlaYap XY X21Yll X22Y21 '- X2aYa/ X21Y12 X22Y22 -- X2aYa2 -- X21Ylp X22Y2p -- X2aYap ... ... ... ...
~XalYll Xa2Y21 ... XanYal XnlY12 Xn2Y22 +--- +XanYa2 -- XnlYlp +Xn2Y2p +~ +XpnYnp~

or -n n n laYal ~ X2nYn2 ~ X2aYnp i~l f~
n n n XY _ ~ X2aYnl ~ X2nYn2 ~ X2aYnP
-- i~l i~l f-l ... ... ... ...
n # n maYnl ~ XmaYn2 .. ~ XmnYnp i=l i=l ~=1 -or even N

Xy,J. = ~Xi~Y,y i~l that is, the calculation for the cell in row i column j of the result matrix, is the sum, for all n, of the products of the nth cell in row i of X with the nth cell of column j of Y.

There are a number of different approaches to inverting a matrix. In the ~lt;fellt;d embodiment of the present invention, the technique used may be shown as:

xx l = adj(X) XI

Here the adjoint of a matrix X (design~t~.d Adj(X)) is divided by the determinant of X (de~ign~ted X). The determinant of any matrix is a single wo 95/24012 2 1 6 2 1 3 1 32- PCT/US95/02557 value, the adjoint of a matrix is another matrix of the same order as the original matrix. Dividing a matrix by a single value is simply dividing each cell of that matrix by the single value.

To calculate the adjoint of a matrix, the determinant of a matrix must be determined. This is achieved by implementing a recursive procedure which is well known to one of ordinary skill in the art.

The sign for any cell as (-l)a+i). The signed minor ( l)~l+n lMijl (where Mij is the matrix which remains when row i and column j is removed) is referred to as the cofactor of the cell and is denoted by c~ij. We can therefore write the c~lc~ tion of the determinant as:

lXI = ~ Xina~V

The adjunct matrix for a Matnx X of order n is c~lclll~tecl as:

11 a2l ... a"l Adj(X~ = al2 a22 -. a,2 ... ... ... ...
.aln a2n ' ann.

That is, it is a transposed matrix of the cofactors.

In the preferred embo-lim~ nt, T st~ti~tics are calculated for each independent variable as a measure of the ~i~nific~nre of that variable to the weather impact model 720. Values of the T st~ti~tics above about 1.5 are ple~,~d. the T st~ti.~tics are calculated as:

Ti = ~i s~

WO9~124012 2 1 6 2 1 3 I PCT/US95/02557 .

Where s = Standard Error of the Regression Vj = ith diagonal element of the matrix (X'X)~'.

The calculation s~ is referred to as the Standard error of the independent variable. The Standard Error of the Regression, s, is c~lcul~t~(l as:

s= ESS
~ n-k where the definitions are as above.

The present invention also considers the P-value, another measure of st~ti.~ti~l ci~nifi~nre, in the deweatherization regression. The P-value is the probability of the F statistic. In the preferred embo-lim~nt it is used as a filter. In other words, if the F-statistic is 10%, there is a 90% prvbability that there is at least one explanatory variable in the weather impact model 720.
In the p~kre,l~d embodiment, the P-value used is an ~plo~ ion rather than a precise calculation which involves solving integrals. This ap~loacll is apparent to one of ordinary skill in the relevant art. The terms N-k and k-l which we used above are referred to in this algorithm as v and u ~t;~ecLi~ely.

Additional ~ c~ ion regarding this and the above st~ti~tie~l approaches may be found in Peizer, D.B. & Pratt, J.W., "A Normal Approximation For Binomial, F, Beta, And Other Common Related Tail Probabilities, " J. Am.
Stat. Assoc. 63:1416-1456 (1968) and Derenzo, S.E., "Approximations for Hand C~lc~ tors Using Small Integer Coefflcient~, " Mathematics of Computation 31:214-222 (1977).

Once the weather impact model 720 has been determined, the correlation l)r~cessor 704 then uses the res-llting weather impact model 720 to forecast dirrelt;ll~ values in step 906. The normal weather data 716 is wo 95n40l2 2 1 6 2 ~ 3 1 PCT~US95/02557 substituted into weather impact model 720 for the historical weather data 714 to arrive at deweatherization data 205. Thus, the weather impact model 720 has to be generated before the deweatherized data 205 can be gener~Led. This is referred to as the deweatherization data 205. Correlation processor 704 then outputs bothl weather impact model 720 and deweatherization data 205.

V. Po~.?~; ,g Processor Figure 10 is a flowchart of the steps performed by forec~ting processor 706. Referring to Figure 10, forec~ting processor 706 receives revised managerial plan 207, weather impact model 720, and deweatherization data 205. Weather impact model 720 and deweatherization data 205 are gene,dt~d by correlation processor 704. The revised managerial plan 207 iS
the original m~n~gerial plan 130 morlifi~d according to the deweatherized data 205.

First, in step 1002, forec~ting processor 706 defines a weatherization regression structure file. The weatherization structure file is similar to the deweatherization r~,t;s~ion structure file rli~c~lssed above. However, the weatherization regression structure file in~hl-les a forecast columns section and a forecast variables section in addition to the four previously mentioned sections under the ~l~t~b~e initi~li7~r.

The forec~ting plocessor 706 then identi~les we~thPri7~tion input files in step 1004. In this step, forecasted weather data 716 iS available from Strategic Weather Services, Wayne, Pennsylvania, U.S.A. to substitute into the weather-impact model 720 in addition to exi~ting deweatherized data 205 and other external inro~ll,a~ion 136.

The forec~ting processor 706 then ~xec~-tes the ~w~all,e,ization regression step 1006. The output of this final step produces the weather-WO 9S/24012 2 1 6 2 1 3 1 PCT/US9~/02S57 .

modified managerial plan 204 which is output to the graphical user interface 710 for replt;senLation and viewing.

Generally, colllpuL~r software evolves in layers of program development, with the most basic layer being the collll,uLel code for the various colllpuLel operating systems. Other layers of software incol~ol~Le, typically under commercial license, pre-e~isting software programs as building blocks for innovative sorLw~r~ to extend colllpuL~r functionality.

In the preferred embodiment of the present invention, the l~t~h~e initi~li7er 702, correlation processor 704, forec~tin~ pr~cessor 706, are implementedwithcommerciallyavailablemulti-(iimP.tl~ionalsoftwareproducts, such as the 'ONE-UP' product, developed by Comshare Inc., Ann Arbor, Michig~n The graphical user interface 710 is implemented in the 'COMMANDER' graphical user interface product m~nllf~rtllred by Comshare Inc. However, the present invention is not limited to these products, and conLelllplates any multi-~limen~ional modelling tool or SQL (Structured Query Language) based ~l~t~h~e or graphical user interface ~r~)acl1 with similar or greater functionality.

Although the invention has been described and ill-lCt~t~l with a certain degree of particularity, it is understood that those skilled in the art will recognize a variety of additional applications and appr~liaLe mo-lifi~ti~ns within the spirit of the invention and the scope of the claims.

Claims (36)

What is Claimed is:

1. A computer-based system for generating a weather-modified managerial plan that represents the future impact of weather and other factors on retail sales, comprising:
first means for providing store information;
second means for providing weather data, said weather data including, historical weather data, normal weather data, and forecasted weather data;
third means for providing information external to retailer store environment;
database initializer means for performing one or more transformations of said store information, said weather data, and said external information to produce variable databases;
correlation processor means for generating a weather impact model, said weather impact model expressing a correlation between said store information and said external information contained within said variable databases with said historical weather data, and for substituting said normal weather data for said historical weather data in said weather impact model to generate deweatherized data; and forecasting processor means for substituting said forecasted weather data for said normal weather data in said weather impact model to produce a weather-modified managerial plan.

2. The system of claim 1, wherein said deweatherized data is made available to a retailer to modify one or more portions of said managerial plan, resulting in a revised managerial plan.

3. The system of claim 2, wherein said substitution of said forecasted weather data in said weather impact model for said normal weather data is performed in conjunction with said revised managerial plan.

4. The system of claim 1, wherein said normal weather data comprises a 30 year average of a particular weather parameter for a specific location and time.

5. The system of claim 1, wherein said store information comprising point of sale data generated at one or more store locations and other store data, said other store data comprising any user specified recordableevent taking place in support of consumer sales.

6. The system of claim 5, wherein said other store data comprises inventory transfer from distribution center to store, promotion data, store traffic, and labor data.

7. The system of claim 5, wherein said point of sale data is generated by electronic scanners, registers, and other electronic scanning and data gathering devices.

8. The system of claim 5, wherein said point of sale data includes product data, POS receipt data, promotional data, markdown data.

9. The system of claim 1, wherein the system is resident within a computer-based MIS architecture.

10. The system of claim 9, wherein said MIS architecture includes a data storage and retrieval facility configured to store said external information and said internal information.

11. The system of claim 1, wherein said external information comprises economic, demographic, competitive, or any other retailer-specified information valuable to assessing business performance.

12. The system of claim 1, wherein said weather data is measured in any user defined time increment.

13. The system of claim 12, wherein said time increment comprises daily time increment.

14. The system of claim 12, wherein said time increment comprises weekly time increments.

15. The system of claim 12, wherein said time increment comprises monthly time increments.

16. The system of claim 1, wherein said database initializer maps store locations to selected areas, thereby utilizing only necessary portions of said weather data associated with said selected areas.

17. The system of claim 1, wherein said database initializer creates retailer hierarchial structure tables, said retailer hierarchial structural tables indicating retailer-specified parent-child relationships.

18. The system of claim 1, wherein said weather data comprises any meteorological factor.

19. The system of claim 18, wherein said meteorological factors include temperature, precipitation, humidity.

20. The system of claim 1, wherein said managerial plan comprises buying product, distributing product to specific locations, advertising, promotions, financial budgeting, historical analysis, labor scheduling, and store traffic analysis.

21. The system of claim 1, wherein said forecasted weather data is in user-determined time increments for user-determined locations.

22. The system of claim 1, wherein said substitution of said forecasted weather data in said weather impact model for said normal weather data is performed in conjunction with said managerial plan.

23. The system of claim 1, wherein said substitution of said forecasted weather data in said weather impact model for said normal weather data is performed in conjunction with said deweatherization data.

24. The system of claim 1, further comprising a data manipulator means, coupled between said first means and said database initializer means, for aggregating said store transaction data.

25. The system of claim 1, wherein said first means comprises:
data gathering means for recording said store transaction data; and a data storage and retrieval facility configured to receive said store transaction data from said data gathering means, and configured to store said store transaction data in a computer-readable format.

26. The system of claim 1, wherein said database initializer means transforms an aggregate of said store transaction data to a form that adds valueto said correlation processor means.

27. The system of claim 26, wherein said database initializer means comprises:
comparing means for comparing a first set of values occurring during a first interval in a first period with a second set of values occurring in a second interval in a second period, and for computing the change in said first and second sets of values, wherein said first and second intervals are equal to a first length of time and said first and second periods are equal to a second length of time, and wherein said first length of time is less than or equal to said second length of time.

28. The system of claim 27, wherein said first length of time is a week.

29. The system of claim 28, wherein said second length of time is a year.

30. The system of claim 27, wherein said first length of time is a day.

31. The system of claim 30, wherein said second length of time is a year.

32. The system of claim 27, wherein said database initializer means further comprises:
lagging means for lagging said logged values of said historical weather data by a third interval, a fourth interval, and a fifth interval.

33. The system of claim 1, wherein said correlation processor means is configured to perform a least squares multiple regression on said variable databases to produce said weather impact model.

34. The system of claim 33, wherein said weather impact model expresses a correlation between said store transaction data contained in said variable databases and said historical weather data.

35. The system of claim 1, further comprising a graphical user interface for receiving said weather-modified managerial plan from said application processor means and for displaying said weather modified managerial plan in a user-specified manner.

36. A computer-based system for generating a weather-modified managerial plan that represents the future impact of weather and other factors on retail sales, comprising:
first means for providing store information;
second means for providing weather data, said weather data including, historical weather data, normal weather data, and forecasted weather data;
third means for providing external information;
database initializer means for performing one or more transformations of said store information, said weather data, and said external information to produce variable databases, wherein said database initializer maps store locations to selected areas, thereby initializing only necessary portions of said weather data associated with said selected areas, and further wherein said database initializer creates retailer hierarchial structure tables, said retailer hierarchial structural tables indicating retailer-specified parent-child relationships;
correlation processor means for generating a weather impact model, said weather impact model expressing a correlation between said store information and said external information contained within said variable databases with said historical weather data, and for substituting said normal weather data for said historical weather data in said weather impact model to generate deweatherized data; and forecasting processor means for substituting said forecasted weather data for said normal weather data in said weather impact model to produce a weather-modified managerial plan.