WO 2013/052081
`
`PCT/US2012/000426
`
`[0224]
`
`If the shopper selects “whole-wheat”, the user has identified. a sub-classification
`
`containing commercial substitutes. There are now.a suitably small set (e.g., fewer than 20) of
`brands, sizes and types of pre-packaged whole wheat bread, from which the system can select the
`lowest price product. Oncealevel containing commercial substitutes is reached, the system
`displays an "ADD"button. Optionally the system allows the user to manually select their
`
`favorite SKU by “whole wheat” display as brand choices. If the user makes a manualselection,
`
`the user's preference for this type of whole wheat bread is recordedin the database. The system
`also remindsthe user that more moneycan be savedif the user selects a sub-set descriptor
`
`instead of a specific SKU,and allows the system to makethe selection: If the user does not make
`
`10
`
`a manualselection of a specific SKU,or if the user returns to the "whole wheat bread"
`
`classification subset, the system selects the lowest cost loaf from the set. Specifically, the lowest
`
`cost pre-packaged whole-wheat bread.
`
`[0225]
`
`This Keyword-Category-SKU Funnelallows the shopper Appsto start with a
`
`15
`
`reasonable approachto selecting a specific item for consumers(as outlinedabove) and then
`quickly utilize “learning”to refine the approach overtime. |
`[0226]
`Logic Outline:
`
`[0227]
`
`The system has a database whichis organized into several (e.g., 10-11)
`
`classification levels. The shopper begins by typing in a string or keyword. As a shoppertypes,
`
`20
`
`the shopper App of the system begins a keyword search of meta data, text, or of the available
`
`item database. In some embodiments, the database is organized so thatat level 6, each of the
`
`classification sub-set descriptors defines a set of commercially equivalent, same or similar
`
`products. The sub-set descriptors or SKUsator belowthis level are displayed to the user with
`
`an ADDbutton,indicating that the remaining items within that sub-set are considered
`
`25
`
`commercially equivalent substitutes. so that the system can make an automatic selection from that
`
`sub-set without more lower-level selection by the user. The classifications above level 6 are
`
`arranged so thatit is generally possible to traverse the index from a search term (e.g., bread) to
`
`the level containing commercial substitutes within three or some other small number(e.g., 2)
`
`sub-set descriptor selections. Although the system allowsthe user to continue to drill down to
`
`30
`
`level 0 or the lowest level, the greatest savings potential is provided when the userclicks the
`
`ADDbuttonto add to his or her cart the sub-set descriptor for a broader sub-set.
`
`41
`
`

`

`WO 2013/052081
`
`PCT/US2012/000426
`
`[0228]
`
`Whenthe shopperfirst enters a searchstring, for each specific item returned, the
`
`search algorithm (SA) displays the unique Level 8 outcomes. So if the user searches on “soup”
`and there are items that are described with the term “soup”in three level-8 groups the system
`causes the user's mobile device to display the three level-8 groups For example, the system can
`37 Ge
`
`display, "soup,”
`
`“soup mix,”or “‘soup base.”
`
`[0229]
`Selecting the Level-8 group that corresponds with what the consumer wants(.e.g.,
`soup), the system will cause the MDto display the appropriate level 7 groups that go with the
`level 8 selection. (e.g., Ready-To-Eat, Condensed, Powdered, Mix).
`[0230]
`Selecting the level 7 group that corresponds to what the user wants
`(e.g.,Condensed) will display the level 6 groups that correspond. (e.g, for a selection of "
`
`10
`
`condensed," the level 6 groups of ’chicken noodle”, “beef”, “chicken with rice”, etc. are
`
`displayed,) and ADD.In this example, the level of specificity is low enough,e.g., the Level 6
`
`descriptors are items considered commercial substitutes, for the system to quickly make a
`
`selection, or for the user to make a manualselection.
`
`15
`
`[0231]
`
`' Selecting the level 6 group that corresponds(e.g., chicken noodle) will display the
`
`level 5 groups that correspondto that sub-set descriptor. For example, the sub-set descriptors
`may be "chicken noodle" and "homestyle chicken noodle".
`[0232]
`Selecting the level 5 group will display the Level 4 brands groups that correspond
`
`20
`
`with Level 5 (e.g., Campbell’s, Wegmans, Acme,Progresso).
`[0233]
`The system allowsthe user to optionally continue to define further attributesthat
`can further reduce the solution set. The database is optimized so that the greatest savings can be
`realized ifthe user allows the system to select a product, once the user has made enough
`classification sub-set selections to reach a sub-classification in which all of the candidate items
`
`are commercial substitutes for one another, so that the system can select and recommend the
`
`25
`
`lowest priced remaining candidate item. Generally, the database is organizedsothat this level is
`reached bythe timethe user has drilled down to level6of the index. The consumer can continue
`
`to drill down to Level 1 or the lowest populated level. Selecting a lower level sub-set descriptor
`
`or specific SKU using the ADD buttonafter level 6 will result in a message such as, “We can
`save you moreif you choose[ level 6 name]”. If the user completes the selection and ADDsthe
`lower-level sub-set or specific SKU to the user's cart, the system includes that specific product in
`
`30
`
`the basket at each store, but continues to determine lowest basket cost at each store by automated
`
`42
`
`

`

`WO 2013/052081
`
`PCT/US2012/000426
`
`selection of a recommended product corresponding to each otheritem in the user's list.
`
`(Thatis,
`
`the system recommendsthe lowest price product from amongst the pre-defined commercial
`
`substitutes, unless the user has ADDeda specific SKUto his/her shoppingcart). If a specific
`SKUis added the system includesthat item in the basket. On the other hand,if the user elects a
`
`sub-set descriptor at Level 6, the system addsthat level 6 sub-set descriptionsto the list and will
`
`later optimize on price lookingat all commercial substitutes represented bylevel 6.
`[0234]
`The system permits the user to search on at least one string descriptor and find
`
`any product from theat least one preselected store in less than a predetermined numberofinputs.
`This predetermined numbercan be 10,9, 8, 7, 6, 5, 4,3, 2, or 1.
`
`[0235]
`
`FIG. 25 shows an example in which the user inputs the string "seventh". The
`
`mobile device displays several sub-classifications including the term, "seventh", The system
`
`displays the next level sub-classification descriptors, including "seventh generation household,"
`"seventh generation liquid,” "seventh generation dish," "seventh generation trash," and "seventh
`
`generation unscented,”
`
`15
`
`[0236]
`
`'
`
`If the user selects the next lower sub-classification "seventh generation dish" then
`
`the sub-classification descriptors including that string are displayed, as shown in FIG. 25B. In
`
`the case where the sub-classification has fewer than a predetermined number(e.g., fewer than
`
`20), all of the specific SKU #s are displayed with the ADD button. Thusin FIG. 25B, the user
`
`has reached a smaller sub-set of substitutes than is desired for minimizing price. By entering a
`
`20
`
`specific string, the user has reached a lowerlevel of the index (below level 6).
`[0237]
`In FIG. 25C,if the user uses the ADD button to add one ofthe specific SKU #s to
`
`her shopping cart, the system displays a reminder that the system can potentially save more
`
`moneyfor the customerif the user specifies a category instead of an SKU#. The mobile device
`suggests adding "dish soap" instead of "seventh generation dish liquid, lemongrass & Clementine
`zest." If the user accepts this suggestion, then the item in the shopper's list is changed to "dish
`
`25
`
`soap."
`[0238]
`
`Each successive selection acts as an additive ("Boolean") filter to whatis
`
`displayed. For example, assumethere are 1500 itemslisted with “bread”in the description, with
`
`3 level-8 groups of 500 products each. Once the Level 8 selection is made the search is now
`
`30
`
`only considering the 500 products of the combinedfilter of “bread” and Level 8 group.
`
`43
`
`

`

`WO 2013/052081
`
`PCT/US2012/000426
`
`[0239]
`In some embodiments,allthe sub-set descriptors at a given classification level are
`mutually exclusive, and each remaining productat a classification level correspondsto exactly
`one ofthe sub-setclassifications at that same level.
`In other embodiments, the database can be
`
`organized so that there is some small overlap between different sub-set classifications at that
`
`same level. For example, a rye-whole-wheat loaf may be classified under, rye bread, whole
`
`wheat bread and multi-grain bread. By defining the sub-set descriptors to minimize the number
`of such overlaps, the numberof levels of selection through which the user drills down before a
`reaching a sub-classification containing only commercial substitutes is reduced.
`
`10
`
`[0240]
`In other embodiments, a small number of miscellaneousitems at a given
`classification level may not correspondto any of the sub-set descriptors in that level (in which |
`case, selection of any of the sub-set descriptors at that level or a lower level will eliminate that
`
`small number of miscellaneous items from being considered in the final automated product
`
`selection. For example, an item maybeeligible for selection by the system at level 6, if the user
`
`clicks the ADD button onalevel 6 sub-set descriptor, but excludedif the user makes a selection
`at a lowerlevel.
`
`[0241] In some embodiments, the system uses a Rule ofN — (whereNis a predetermined
`
`input variable, for example, N=20) This rule permits the user to make fewer selectionsif the
`initial search by the user corresponds to a lowerlevel (e.g., level 5 or 4) sub-classification and
`only a few SKUs correspond. The algorithm will follow the steps described above. If the total
`
`20
`
`number of remaining SKUs(after application of the combinations of the applied filters) at any
`
`level below level 6 is <=N (e.g., <=20) items, then the algorithm will display those items for
`
`optional user selection immediately.If the highest level at which the number of remaining items
`(after application of the combinationsof appliedfilters) <=N (€.g., <=20)is level 6 then the
`algorithm will display the level 6 results with the ADD button. Once the user has clicked the
`"ADD"button, the user has provided sufficient information for the system to make an automated
`selection of a lowest price commercial substitute products at each store, corresponding to one of
`the itemsin the user's list. If the user has ADDeda sub-set identifier,
`the system adds the item
`
`25
`
`(with whateverlevel of specificity the user has provided) to the user's basket for cost comparison
`
`between stores amongst similar items.
`
`30
`
`[0242]
`
`The rule of N does not continue beyond level 6. The shopper can continueto drill
`
`down andspecify additional sub-set descriptors, which act as filters, but the rule will no longer
`
`44
`
`

`

`WO 2013/052081
`
`PCT/US2012/000426
`
`be used. E.g., Suppose the selection of “Banana Bread” results in 4-SKUs remaining at a
`retailer, the algorithm will display Banana Bread-Loaf, Banana Bread-Mix, Banana Bread-
`
`Muffins . The routine follows the steps outlined above but does NOT makethe userclick at
`Levels 10, ...., 1. Rather, the system will display the lowestlevel (higher of 6 and 5) with ADD
`buttons immediately.
`.
`[0243]
`In some embodiments,if the results of the initial search <= 20 (at Level 6) then
`
`the items will be displayed with description of Level 6+Level 9. E.g., if the user searches,
`
`“Whole Wheat Tortilla" the search returns and displays “Whole Wheat Tortilla” (redundant
`
`Tortilla is eliminated).
`
`,
`
`Search Termsare additive in a Boolean sense. E.g., Whole wheat Tortilla
`[0244]
`searchesfor items with the terms “Whole AND Wheat ANDTortilla” will be matched against
`
`Level 3 — Level 10 descriptors, high level item description, detailed item descriptions
`[0245]
`In some embodiments the algorithm will do an OR search of multiple keywords
`entered at levels 10 through 8 AND Level 6-3 . If there is a match at level 8 AND a matchat
`
`Levels 6-3 the system will display subset of the identified level 8 categories which also match at
`
`level 6. If there is no matchat levels 6-3 then the matching level 8 categories will be displayed;
`
`a standard (single keyword) drill down will start and the remaining rules of a simple search
`
`outlined herein will follow for the remainder of the search.
`
`20
`
`[0246]
`
`[0247]
`
`Price Minimization Algorithm (MAP)Performance
`
`Once the user has selected a respective item (sub-set identifier or SKU) for each
`
`item in the user's list, the system uses a price minimization algorithm,to find a local minimum
`
`price thatsatisfies the user's list. The result is a referred to herein as a "minimum,"in that the
`total set of possible combinationsis initially trimmedto reflect the user's selection of stores and
`
`25
`
`preferencesettings. It is not necessarily the absolute minimum amongevery productsold at any
`
`store. Further, in some embodiments, the total set of products in the database which are
`
`considered as substitutes is trimmedto include a predetermined percentage (e.g., 80%) of the
`
`products of a given type. For example, stores may carry 30 different types (brands) of canned
`
`10.5 oz cans of condensed chicken noodle soup. If 80% of all sales of canned 10.5 oz cans of
`
`30
`
`condensed chicken noodle soup are made up ofonly the top ten sellers out of the 30 products,
`
`then the database can be trimmedto exclude the 20 products which make up the remaining 20%
`
`45
`
`

`

`WO 2013/052081
`
`PCT/US2012/000426
`
`of total sales. Although oneof these excluded products may have the absolute minimum price,
`if sales are used as a proxy for quality and preference,then the top selling products making up
`
`the 80% oftotal sales are viewed as having comparable, acceptable quality or as preferred by
`
`most shoppers. Thus, the top sellers are considered by the system as being substitutes for each
`other. The minimum price determined by the system is not an absolute global minimum price,
`
`but a minimumprice ofa basket of products satisfying the user's preferences and meeting a level
`
`of market share or quality established by the administrator of the system database or users. The
`
`user's preferences. further allow the user to select a higher quality standard, so that only the
`
`products identified as "super-premium products" by the administrator are consideredfor this
`
`10
`
`particular user,if there is a premium product version of one ofthe items on the user'slist.
`
`[0248]
`
`[0249]
`
`Setup/Structure:
`
`1.
`
`In some embodiments, the system assignsan initial default set of preferences
`
`to each user. In some embodiments, the system providesthe user with series of options for
`preferences, and allows a newuserto self-identify his or her own preferences or affinity
`group(s). For example, in a grocery store setting, the system can allow the shopperto select a
`preference for super-premium food products. If the user selects this option, then the set of
`candidate products suggested to the user is reduced to exclude lower priced / lower quality items.
`This may be done by ranking the products according to price and selecting those products priced
`abovea cutoffprice as candidates, or may be done by manually selecting the group to include in —
`(or exclude from) the candidate set, or it may be donebythe user, or itmay be done by crowd
`source opinion of what products are “super premium”. In another grocery store example, the
`
`system can allow the shopperto identify himself/herself as an extreme bargain hunter, in which
`
`case all products in the database (possibly including some products of lower quality and large
`
`sized units with lower price per ounce) are includedin the initial candidate sets for that user.
`
`In
`
`this manner the system can establish a variety of product preference “presets” that represent
`different approachesto considering products as commercialsubstitutes and the shopper can
`select one of these presets as an initial starting point for their system preferences. Once the
`shopper begins using the app the system will optionally modify these initial setting to match the
`Shopper’s actual behavior. The system creates for each shoppera database of product
`substitutes (SdBPS) that will have the following initial presets-
`
`15
`
`20
`
`25
`
`30
`
`46
`
`

`

`WO 2013/052081
`
`PCT/US2012/000426
`
`[0250]
`
`a.
`
`For each product category the system will pre-define a numberof default top
`
`In the
`brands(e.g., the top 4 brands) or products that make up the majority of sales volume.
`Preference Engine these4 brands will be considered perfect substitutes for each other. For
`example: Category= Spaghetti Sauce, Brand 1=Prego, Brand 2=Ragu, Brand 3=Francesco
`
`Rinaldi, Brand 4= Classico's. The system receives as inputs sales data for various products,
`
`which the system sorts by sales volume, and from which the system automatically identifies
`
`these top sellers, which are stored in the database. Alternatively, the System Administrator
`
`determines which product makes up the bulk of user demand.
`
`[0251]
`
`b.
`
`For each category the system will pre-define a main productsize or package
`
`10
`
`that represents the majority of sales on a size or pack basis. For example: Category = Spaghetti
`
`Sauce, Main product size = 240z. In some embodiments, the system correlates demographic
`
`information with size (e.g., households having four or more people most often choose the 32 oz.
`
`size, but houscholds having 1-3 people most often choose the 16 oz. size). This information can
`
`be used in conjunction with the information obtained from the user during registration, to set the
`user's pre —defined size, via pre-sets or other setup capabilities.
`
`15
`
`[0252]
`
`c.
`
`For each Brand-size combination the system will identify specific flavors or
`
`items that are similar. For example: Category = Spaghetti Sauce, Flavor = Tomato Basil is the
`
`same as Tomato & Basil and Tomato and Basil.
`
`20
`
`25
`
`[0253]
`2.
`Shoppers will have previously selected up to a predetermined number(for
`example, but not limited to, 4) of stores that the shopper considers acceptable substitutes from a
`shopperperspective. That is, these stores satisfy the shopper's criteria for convenience and
`
`In addition the shopper will have created an initial shoppinglist of itemsat the
`perceived value.
`categories and/or specific item level.
`
`.
`
`3. The system providesa real-time database of individual products including
`[0254]
`a. Categorization of each item’at brand, sub-brand,size, flavor, pack and other
`[0255]
`descriptor levels (e.g., Organic, hormonefree etc) -. The categorization accordingto levels will
`be based on a fixed or “dynamic”structure for different product categories,e.g., thereby
`
`effectively below 10 levels.
`
`[0256]
`
`b.
`
`For each item andstore/price zone, the system will also keep a table or
`
`30
`
`database of shelf prices and deadnetprice (the price of an item incorporating all promotional
`
`47
`
`

`

`WO 2013/052081
`
`PCT/US2012/000426
`
`discounts of any kind including coupons, temporary price reductions, shoppercard discounts
`etc.)
`
`For each Category and each item in a category and store/price zone,the
`c.
`[0257]
`system will calculate the comparable price per unit (e.g., deadnet price/oz).
`
`[0258] 4. Shoppers typically makealist of “category” descriptors — e.g., “Soup.”
`
`Shoppers know whatthey consider suitable matches for the descriptor “Soup.”
`
`[0259] _5. Whenashoppertypes in a category item — e.g., “Soup” the Appofthe
`
`system performs a matchingroutine that will identify the specific item at a low enoughlevel in
`
`the data architecture (Level 6 or Level 5 in the Table 1 example). The matching functions as
`
`10
`
`follows:
`
`[0260]
`
`a. An item addedto thelist is keyword matched to the highest database product
`
`level, corresponding to the descriptor, in the database ofall grocery store items. Here, “Soup”is
`
`matchedto level 9 and 8 sub-set descriptors, but it appears uniquely in level 8. Thus, the
`
`algorithm then returns Level 7 descriptors for selection. E.g., “Condensed, Ready-to-Eat, Mix,
`
`15
`
`Microwaveable”
`The shopper then selects one ofthe Level 7 descriptors. E.g., The shopper
`[0261]
`b.
`selects “Ready-to-Eat.”
`
`[0262]
`
`c. Once the shopperhasselected the Level 7 descriptor, the app now treats the
`
`selected descriptors as additive filters (e.g.,: in Boolean terms, “Soup” AND “Ready-to-Eat”) and
`
`20
`
`returns the descriptors of Level 6 that remain from this search.
`
`[0263]
`
`d.
`
`The App displays Level 6 descriptors which represent in Table 1 “flavors.”
`
`Selection of a Level 6 descriptoris sufficient for the App to make pricing
`€.
`[0264]
`choice decisions and so the App displays an “Add”button nextto the flavors displayed in Level
`
`6. E.g.,. App displays “Chicken Noodle, Tomato, New England Clam Chowder.” The shopper
`
`25
`
`selects “New England Clam Chowder”.
`
`[0265]
`
`f.
`
`Ifthe shopper would like to drill down deeper and be more specific the app
`
`allows the shopperto tap on the displayed item to moveto levels 5, 4, 3, 2, 1, 0 filters with the
`
`“Add”button displayed at each level. The system does not alwaysuse all 10 levels. The string
`
`or keywordinitially entered by the user may correspond to a lowerlevel sub-classification
`
`30
`
`descriptor. Further, the user may choose the ADD buttonat one ofthe intermediate levels (e.g.,
`
`level 6), so that the system can choosethe least expensive product at an intermediate level.
`
`48
`
`

`

`WO 2013/052081
`
`PCT/US2012/000426
`
`[0266]
`
`g.
`
`Ifthe user makes a selection at levels lower than necessary (for the system to
`
`automatically identify a set of equivalent products and recommenda lowest cost product among
`
`them) the system provides a warning message explaining to shoppers that more money can be
`saved by allowing the app to maketradeoffs at higher(less specific) levels within the database.
`
`[0267]
`
`6. When the “SaveOn!”function is selected and run items on the shopperslist
`
`(Pre-Algorithm List) may be described by the Shopperin degrees of specificity ranging from
`
`broad to specific. All items on the list are described at a level that is at least as specific as to
`identify commercial substitutes within the product category. All items on the list are specified at
`
`most-at the SKU level. For each item the MAP will look at related items (either related
`
`10
`
`commercial substitutes, a more specific level of categorization but no more than SKU). For each
`
`item the MAP will compare the dead-netprices/unit ofall substitutable items, if any, for each
`
`category,identifying the lowest priced item (price/unit) for the first store of the up to the
`
`predetermined number(e.g., 4) of stores selected above.
`[0268]
`7.
`The MAPproceeds to make the same comparison for each item on the Pre-
`Algorithm list, until the lowest priced item is identified for each item (as now identified by the
`selected classification and sub-set descriptors) on the Pre-Algorithm Listat the first store.
`[0269]
`8.
`The MAPthen conducts the same type of comparison for the same Pre-
`Algorithm list of items at each ofthe selected stores — Store 1, Store 2, Store 3 and Store4...
`Store N). Note that the basket of products selected at each store, correspondingto the pre-
`algorithm list, can contain a different set of SKUs for each respective store. The system takes
`
`20
`
`into accountthe actual prices at each store.
`[0270]
`9. When the lowestpriced basketis identified for each of the selected stores —
`the basket costs are then comparedacross stores and the lowest overall basket and store
`
`combination (the Post-Algorithm List) is identified and presented to the shopper.
`
`25
`
`[0271]
`
`10.
`
`In certain cases the app mayfirst calculate the total basket cost at a retailer so
`
`that retailer couponsthat are based on a specific basket cost target can be reflected prior to
`inclusion of other couponsorsales that might put the basket below the targeted amount.
`[0272]
`11. The MAPalso identifies the highest priced basket-store combination looking
`at the same size, category combinations identified for the lowest priced basket for each store.
`
`30
`
`Thus, in this computation the same basket of products which is the lowest amongall of the stores
`
`49
`
`

`

`WO 2013/052081
`
`PCT/US2012/000426
`
`is then priced at all of the stores. The result is an "apples-to-apples" comparison for a basket of
`specific products (brands, sizes and flavors) at each of the selected stores.
`[0273]
`12. When the lowest, and highest, priced basket-store combination has been
`identified,
`the system displays the lowestpriced Post-Algorithm List along with the nameofthe
`target store(i.e., the store having the lowesttotal price for the basket). The savings level and
`total basket price foreachstoreis also provided within the app for consumers tapping on display
`of “winning”store results.
`[0274]
`13. When the Post-Algorithm list is displayed (specific items) the following are
`
`also displayed:
`[0275]
`
`a.
`
`10
`
`For each item on the recommendedlist, icons showing whattypes of
`
`discounts were incorporatedinto the final price are displayed next to the item (e.g., an icon for
`coupons, an icon for price reductions, and Icon for shopper card promotions)
`[0276]
`b. The baskettotal cost is displayed along with the $ and % savings versus the
`
`highest priced basket-store combination identified above in #8. In some embodiments,the
`
`15
`
`savings are calculated relative to different reference points.
`
`[0277]
`
`14. The system then looks to the web-based direct marketing Campaign
`
`Management System (or other reference points) to see if any offers are available on the items in
`the Post-Algorithm List and store recommendation. If there are available offers, the system will
`signal the shopper that additional savingsare available. The signal- will be clearly associated
`
`20
`
`with the relevant Post-Algorithm List items.
`
`[0278]
`
`15. When a shopperselects an item that has an additional offer associated withit,
`
`the offer details are presented (e.g., Shelf price, $/off, deadnet price, purchase requirements and
`
`brand/item details). Shoppers can acceptor reject the incremental promotionaloffer.
`
`“Accepting” the extra offer will cause the Post-Algorithm list to be updated to reflect any
`
`25
`
`changesin brand, size, purchase quantity, flavor necessary to accept the offer. The total basket
`
`price is updated and displayed along with updated $ and % savings versus the Highest-Priced
`basket with the same change madein the highest-priced basket (but without any discounts) by
`the selection of the promoted item.
`
`[0279]
`
`16. Once the Post-Algorithm List is presented, the shopper will be allowed to
`
`30
`
`change recommendedproducts by tapping an item and then responding/navigating through a
`
`series of filters to the desired item.
`
`50
`
`

`

`WO 2013/052081
`
`PCT/US2012/000426
`
`[0280]
`
`17. When a shopper changes a recommended product on the Post-Algorithm
`
`List, the SdBPS from #1 above will be updated as follows:
`
`[0281]
`
`a.
`
`The selection of a Brand not previously flagged as a substitute in the SdBPS,
`
`will cause a flag to be set for that item as a suitable substitute in the SdBPS. The flag for the
`
`brand being de-selected will be set to the “off” position or weighted differently in any preference
`
`algorithm. Brandsnot flagged as substitutes will not be compared in the MAPand will not
`
`appearin a post-algorithm list until flagged.
`
`[0282]
`
`b. The selection of a different size not previously flagged as a substitute in the
`
`SdBPS,will cause a flag to be set for that size as a suitable substitute in the SdBPS. The app
`
`10
`
`will dynamically monitor size selection within a category and after a variable numberof
`
`“deselections”the size deselected “N” times will be set to the “Off” position. Sizes not flagged
`
`as substitutes will not be compared in the MAPandwill not appear in a post-algorithm list until
`
`flagged.
`
`15
`
`[0283]
`c.
`The selection of a Flavor not previously flagged as a substitute in the SdBPS,
`will cause a flag to be set for that item as a suitable substitute in the S€BPS.
`The flag for the
`|
`flavor being de-selected will NOTbeset to the “off” position. Flavors not flagged as substitutes
`
`will not be comparedin the MAPandwill not appear in a post-algorithm list until flagged.
`
`[0284]
`
`Special Cases and Situations
`
`20
`
`Private Label Products: Private label, store brand and generic products may
`1.
`_
`[0285]
`be automatically selected as suitable substitutes within the SdBPS. The system will determine in
`
`the initial presets which productsare top sellers and suitably close in quality as to be considered
`suitable substitutes. Products that are of significantly low or lower quality than the top selling
`items in a category will not be flagged as substitutes. “Top selling” is to mean those products
`
`and brands making up ~80% ofa category’s sales and for which general consumerpreference is
`positive. | The system and the Appwill seek to avoid recommending products that may be the
`cheapest — but have by individual consumerchoice in the market — been deemed unacceptable or
`
`of significantly lower quality.
`[0286]
`a. Within the systemsetup screen on the app/website a global function will
`allow the user to override the Private label approach described above and include or exclude
`
`ALLprivate label products or other products from their personal SdBPS and the MAP.
`
`25
`
`30
`
`31
`
`