IJSZOI40379464A1 (19) United States (12) Patent Application Publication (10) Pub. No.: US 2014/0379464 A1 Aronowich et al. (43) Pub. Date: Dec. 25, 2014 (54) BUDGET DISTRIBUTION IN ONLINE (52) US. Cl. ADVERTISING CPC ................................ .. G06Q 30/0249 (2013.01) USPC ..................................................... .. 705/14.48 (71) Applicant: Kenshoo Ltd., Tel Aviv (IL) (72) Inventors: Michael AronoWich, Haifa (IL); Arriel Johan Benis, Rehovot (IL); Reut Yanai, (57) ABSTRACT Tel Aviv (IL); Gil Vind, Hod Hasharon (IL) A method for budget distribution in online advertising, the (73) Assignee: Kenshoo Ltd_ Tel AViV (IL) method comprising using at least one hardware processor for: 3 receiving a de?nition of a single advertiser budget to be spent (21) App1_ NO_; 14/314,151 on advertising multiple ad entities in an online advertising platform; receiving historical performance data associated (22) Filed: Jun. 25, 2014 with the multiple ad entities, wherein the historical perfor _ _ mance data comprises multiple proportional performance Related U-s-APPllcatlon Data metrics for each of the multiple ad entities; computing a (60) provisional application NO_ 61/839,139, ?led on Jun health index for each of the multiple ad entities, the health 25’ 2013_ index being a weighted average of multiple components com prising the multiple proportional performance metrics, Publication Classi?cation wherein the multiple components are each monotonic with respect to spend; and proportionally distributing the single (51) Int. Cl. advertiser budget between the multiple ad entities, based on G06Q 30/02 (2006.01) the health indices of the multiple ad entities. /1 0 12K COMPUTER SYSTEM SERVER 2/8 30 MEMORY 34 RAM ‘— 1 6 STORAG E H SYSTEM PROCESSING CACHE _L 40 UNIT / 32 42/ 181. 'T 24 22 \ \ 20 O / V DISPLAY NETW FlK ADAPTER |NTERFACE(S) O 1 4 \ EXTERNAL DEVICE(S) Patent Application Publication Dec. 25, 2014 Sheet 1 0f 5 US 2014/0379464 A1 f10 12\ COMPUTER SYSTEM SERVER 28 30 MEMORY 34 \ fg RAM <— 16 STORAGE \ SYSTEM PROCESSING CACHE k UNIT 3; 40\‘ / 18x 42 | 24 22 \ \ 20 |/O / DISPLAY INTERFACE (S) NETWORK ADAPTER 14 \ EXTERNAL DEVICE(S) FIG. 1 Patent Application Publication Dec. 25, 2014 Sheet 2 0f 5 US 2014/0379464 A1 [54C / Patent Application Publication Dec. 25, 2014 Sheet 3 0f 5 US 2014/0379464 A1 Software Mapping an Deg/?g)me ClasVisrrtouaolm Data Ari-8W“, Transaction Navigatio Lifec cle Education Processmg Processing y Delivery Manageme Workloads 66 R esource Meteri. ng Servr. ce Level SLA Plannr. n Provrsronlng and Pricing User Portal Management and Fulfillmen Management “F?E@Ei11 Virtual Virtual Virtual Virtual Virtual Servers Storage Networks Applications Clients Virtualization éUWee%@ Mainframes RISC Dalabase Network Storage Networking Architecture Software Application Servers Server Software Hardware and Software 60 FIG. 3 Patent Application Publication Dec. 25, 2014 Sheet 4 0f 5 US 2014/0379464 A1 / 400 RECEIVE DEFINITION OF SINGLE \/\ ADVERTISER BUDGET 402 MONETARY TIME AMOUNT V RECEIVE HISTORICAL PERFORMANCE \A V DATA OF PORTFOLIO MEMBERS 404 V COMPUTE HEALTH INDICES FOR PORTFOLIO MEMBERS 406 BAYESIAN ESTIMATION OF COMPONENTS 406a STANDARDIZATION OF COMPONENTS 406b COMBINE COMPONENTS INTO HEALTH INDEX 4060 V TRANSFORM HEALTH INDICES USING SIGMOID CURVE I 408 V DISTRIBUTE BUDGET BETWEEN PORTFOLIO MEMBERS I 410 FIG. 4 Patent Application Publication Dec. 25, 2014 Sheet 5 0f 5 US 2014/0379464 A1 502 500 0 HEALTH INDEX FIG. 5 US 2014/0379464 A1 Dec. 25, 2014 BUDGET DISTRIBUTION IN ONLINE tion as well as one or more ads. Key advantages of the Inter ADVERTISING net, relative to other information media, are that each page can be customized to ?t a customer pro?le and ads can contain CROSS-REFERENCE TO RELATED links to other Internet pages. Thus, ads can be directly tar APPLICATIONS geted at different customer segments. For example, ad target ing is nowadays possible based on the geographic location of [0001] This application claims the bene?t of US. Provi the advertiser and/ or the customer, the past navigation path of sional PatentApplicationNo. 61/839,139, ?led Jun. 25, 2014, the customer outside or within the web site, the language used which is incorporated herein by reference in its entirety. by the visitor’s web browser, the purchase history on a web site, the behavioral intent in?uenced by the user’s action on FIELD OF THE INVENTION the site, and more. [0002] Present embodiments relate to the ?eld of online [0008] Furthermore, the ads themselves are often designed advertising. and positioned to form direct connections to well-designed Internet pages. The concept referred to as “native advertising” BACKGROUND offers ads which more naturally blend into a page’ s design, in cases where advertiser’ s intent is to make the paid advertising [0003] Advertising using traditional media, such as televi feel less intrusive and, therefore, increase the likelihood users sion, radio, newspapers and magaZines, is well known. will click on it. Unfortunately, even when armed with demographic studies [0009] The foregoing examples of the related art and limi and entirely reasonable assumptions about the typical audi tations related therewith are intended to be illustrative and not ence of various media outlets, advertisers recognize that exclusive. Other limitations of the related art will become much of their advertising budget is oftentimes simply wasted. apparent to those of skill in the art upon a reading of the Moreover, it is very dif?cult to identify and eliminate such speci?cation and a study of the ?gures. waste. [0004] Recently, advertising over more interactive media SUMMARY has become popular. For example, as the number of people using the Internet has exploded, advertisers have come to [0010] The following embodiments and aspects thereof are appreciate media and services offered over the Internet as a described and illustrated in conjunction with systems, tools potentially powerful way to advertise. and methods which are meant to be exemplary and illustra [0005] Interactive advertising provides opportunities for tive, not limiting in scope. advertisers to target their advertisements (also “ads”) to a [0011] There is provided, in accordance with an embodi receptive audience. That is, targeted ads are more likely to be ment, a method for budget distribution in online advertising, useful to end users since the ads may be relevant to a need the method comprising using at least one hardware processor inferred from some user activity (e.g., relevant to a user’s for: receiving a de?nition of a single advertiser budget to be search query to a search engine, relevant to content in a spent on advertising multiple ad entities in an online adver document requested by the user, etc.). Query keyword target tising platform; receiving historical performance data associ ing has been used by search engines to deliver relevant ads. ated with the multiple ad entities, wherein the historical per For example, the AdWords advertising system by Google Inc. formance data comprises multiple proportional performance of Mountain View, Calif., delivers ads targeted to keywords metrics for each of the multiple ad entities; computing a from search queries. Similarly, content-targeted ad delivery health index for each of the multiple ad entities, the health systems have been proposed. For example, US. Pat. No. index being a weighted average of multiple components com 7,716,161 to Dean et al. and US. Pat. No. 7,136,875 to prising the multiple proportional performance metrics, Anderson et al. describe methods and apparatuses for serving wherein the multiple components are each monotonic with ads relevant to the content of a document, such as a web page. respect to spend; and proportionally distributing the single Content-targeted ad delivery systems, such as the AdSense advertiser budget between the multiple ad entities, based on advertising system by Google for example, have been used to the health indices of the multiple ad entities. serve ads on web pages. [0012] There is further provided, in accordance with an [0006] AdSense is part of what is often called advertise embodiment, a computer program product for budget distri ment syndication, which allows advertisers to extend their bution in online advertising, the computer program product marketing reach by distributing advertisements to additional comprising a non-transitory computer-readable storage partners. For example, third party online publishers can place medium having program code embodied therewith, the pro an advertiser’s text or image advertisements on web pages gram code executable by at least one hardware processor for: that have content related to the advertisement. This is often receiving a de?nition of a single advertiser budget to be spent referred to as “contextual advertising”. As the users are likely on advertising multiple ad entities in an online advertising interested in the particular content on the publisher web page, platform; receiving historical performance data associated they are also likely to be interested in the product or service with the multiple ad entities, wherein the historical perfor featured in the advertisement. Accordingly, such targeted mance data comprises multiple proportional performance advertisement placement can help drive online customers to metrics for each of the multiple ad entities; computing a the advertiser’s website. health index for each of the multiple ad entities, the health [0007] Optimal ad placement has become a critical com index being a weighted average of multiple components com petitive advantage in the Internet advertising business. Con prising the multiple proportional performance metrics, sumers are spending an ever-increasing amount of time wherein the multiple components are each monotonic with online, looking for information. The information, provided respect to spend; and proportionally distributing the single by Internet content providers, is viewed on a page-by-page advertiser budget between the multiple ad entities, based on basis. Each page can contain written and graphical informa the health indices of the multiple ad entities. US 2014/0379464 A1 Dec. 25, 2014 [0013] In some embodiments, the multiple components [0034] FIG. 5 shows a sigmoid curve on a Cartesian coor further comprise an absolute parameter. dinate system. [0014] In some embodiments, the absolute parameter is a current reach value for each of the multiple ad entities. DETAILED DESCRIPTION [0015] In some embodiments, the computing of the health index further comprises a Bayesian estimation of at least Glossary some of the multiple proportional performance metrics. [0016] In some embodiments, the computing of the health [0035] “Online advertising platform” (or simply “advertis index further comprises standardizing the multiple compo ing platform”): This term, as referred to herein, may relate to nents. a service offered by an advertising business to different adver [0017] In some embodiments, the health indices are sig tisers. In the course of this service, the advertising business moid-transformed health indices. serves ads, on behalf of the advertisers, to Internet users. Each [0018] In some embodiments, the proportionally distribut advertising platform usually services a large number of ing comprises solving the optimization program: advertisers, who compete on advertising resources available through the platform. The competition is oftentimes carried out by conducting some form of an auction, where advertisers bid on advertising resources. The ads may be displayed (and/ or otherwise presented) in various web sites which are af?li ated with the advertising business (these web sites constitut subject to Z (xk) = B ing what is often referred to as a “display network”) and/ or in one or more web sites operated directly by the advertising [bk < xk < bib, business. To aid advertisers in neatly organizing their ads, advertising platforms often allow grouping individual ads in [0019] where: sets, such as the “AdGroups” feature in Google AdWords (a [0020] k is one ad entity of the multiple ad entities, service operated by Google, Inc. of Mountain View, Calif.). [0021] hk is one health index of the of the kth ad entity, The advertiser may decide on the logic behind such grouping, [0022] B is the single advertiser budget, but it is common to have ads grouped by similar ad copies, [0023] xk represents an unknown budget ration per the kth similar targeting, etc. Advertising platforms may allow an ad entity, and even more abstract way to group ads; this is often called a “campaign”. A campaign usually includes multiple sets of [0024] xk is constrained by lb k and ubk, which are lower and upper bounds, respectively, imposed on the kth ad entity of the ads, with each set including multiple ads. An advertiser may budget ration xk. control the cost it spends on online advertising by assigning a [0025] In some embodiments, the computing of the health budget per individual ad, a group of ads or the like. The budget index further comprises endowing different ones of the mul may be de?ned for a certain period of time. tiple ad entities with different weights, based on business [0036] “Search advertising platform”: A type of advertising logic provided by an advertiser. platform in which ads are served to Internet users responsive [0026] In some embodiments, the multiple components to search engine queries executed by the users. The ads are comprise the click-through rate, the conversion rate, the typically displayed alongside the results of the search engine potential reach, the spend rate and the reach. query. AdWords is a prominent example of a search advertis [0027] In some embodiments, the single advertiser budget ing platform. In AdWords, advertisers can choose between is de?ned for a period of time over which the single advertiser displaying their ads in a display network and/ or in Google’s budget is to be spent; and the receiving of the historical own search engine; the former involves the subscription of performance data, the computing of the health index and the web site operators (often called “publishers”) to Google’s proportionally distributing are performed multiple times dur AdSense program, whereas the latter, often referred to as ing the period of time for which the single advertiser budget SEM (Search Engine Marketing), involves triggering the dis is de?ned, thereby optimizing a spend rate of the single adver playing of ads based on keywords entered by users in the tiser budget during the period of time. search engine. [0028] In addition to the exemplary aspects and embodi [0037] “Social advertising platform”: A further type of ments described above, further aspects and embodiments will advertising platforms, commonly referred to as a “social” become apparent by reference to the ?gures and by study of advertising platform, involves the displaying of ads to users the following detailed description. of online social networks. An online social network is often de?ned as a set of dyadic connections between persons and/or BRIEF DESCRIPTION OF THE FIGURES organizations, enabling these entities to communicate over the Internet. In social advertising, both the advertisers and the [0029] Exemplary embodiments are illustrated in refer users enjoy the fact that the displayed ads can be highly enced ?gures. The ?gures are listed below: tailored to the users viewing them. This feature is enabled by [0030] FIG. 1 shows a schematic of an example of a cloud way of analyzing various demographics and/or other param computing node; eters of the users (jointly referred to as “targeting criteria”)i [0031] FIG. 2 shows an illustrative cloud computing envi parameters which are readily available in many advertising ronment; platforms of social networks and are usually provided by the [0032] FIG. 3 shows a set of functional abstraction layers users themselves. Facebook Ads, operated by Facebook, Inc. provided by the cloud computing environment of FIG. 2; of Menlo Park, Calif., is such an advertising platform. Linke [0033] FIG. 4 shows a ?ow chart of an exemplary method dIn Ads, by LinkedIn Corporation of Mountain View, Calif., for budget distribution in online advertising; and is another. US 2014/0379464 A1 Dec. 25, 2014 [0038] “Online ad entity” (or simply “ad entity”): This [0049] “Conversion rate (CR)”: the total number of con term, as referred to herein, may relate to an individual ad, or, versions divided by the total number of clicks; alternatively, to a set of individual ads, run by an advertising [0050] “Return on investment (ROI)” or “Return on platform. An individual ad, as referred to herein, may include advertising spending (ROAS)”: the ratio between the an ad copy, which is the text, graphics and/or other media to amount of revenue generated as a result of online adver be served (displayed and/ or otherwise presented) to users. In tising, and the amount of investment in those online addition, an individual ad may include and/or be associated advertising efforts. Namelyirevenue divided by with a set of parameters, such as searched keywords to target, expenses; geographies to target, demographics to target, a bid for utili [0051] “Revenue per click”: the average amount of rev zation of advertising resources of the advertising platform, enue generated to the advertiser per click (or another and/or the like. Sometimes, the bid may set for a particular interaction with an ad entity), calculated by dividing parameter instead of or in addition to setting a global bid for total revenue by total clicks; the ad entity; for example, a bid may be per keyword, geog [0052] “Revenue per impression”: the average amount of raphy, etc. revenue generated to the advertiserper impression of the [0039] “Reach”: the number of users which ?t certain tar ad entity, calculated by dividing total revenue by total geting criteria of an ad entity. This is the number of users to impressions; which that ad entity can be potentially displayed. The “reach” [0053] “Revenue per conversion”: the average amount of metric is common in social advertising platforms, such as revenue generated to the advertiser per conversion, cal Facebook. culated by dividing total revenue by total conversions; [0040] “Search volume”: the number of average monthly [0054] “Unique-impressions-to-reach ratio”: the ratio searches (or searches over another period of time) for a cer between the number of unique impressions (i.e. impres tain search term. The search volume is often provided by sions by different users, ignoring repeated impressions search advertising platforms, such as Google AdWords. by the same user) and the reach of the ad entity. This ratio [0041] “Performance”: This term, as referred to herein with represents the realized portion of the reach. regard to an ad, may relate to various statistics gathered in the [0055] “Spend rate”: the percentage of utilized budget course of running the ad. A “running” phase of the ad may per a certain time period (e.g. a day) for which the budget refer to a duration in which the ad was served to users, or at was de?ned. In many scenarios, even if an advertiser least to a duration during which the advertiser de?ned that the assigns a certain budget for a certain period of time, not ad should be served. The term “performance” may also relate the entire budget is consumed during that period. The to an aggregate of various statistics gathered for a set of ads, spend rate metric measures this phenomenon. a campaign, etc. The statistics may include multiple param [0056] “Quality score”: a score often provided by adver eters (also “performance metrics”). Exemplary performance tising platforms for each ad entity. For example, Google metrics are: AdWords assigns a quality score between 1 and 10 to [0042] “Impressions”: the number of times the ad has each individual ad. Factors which determine the quality been served to users during a given time period (e.g. a score include, for example, CTR, ad copy relevance, day, an hour, etc.); landing page quality and/or other factors. The quality [0043] “Frequency”: the average number of times a user score, together with the bids placed by the advertiser, are has been exposed to the same ad, calculated as the ratio usually the factors which affect the results of the com of total number of impressions to the number of unique petition between different advertisers on advertising impressions (i.e. the number of unique users exposed to resources. that ad). This metric is very common in social advertis [0057] “Potential reach”: de?ned as 1 minus the unique ing platforms; impressions-to-reach ratio. The higher the potential [0044] “Clicks”: the number of times users clicked (or reach, the more users are left to display the ad entity to. otherwise interacted with) the ad entity during a given [0058] “Proportional performance metrics”: those of the time period (e.g. a day, an hour, etc.); above performance metrics (or other performance metrics not [0045] “Cost per click (CPC)”: the average cost of a click discussed here) which denote a proportion between two per (or another interaction with an ad entity) to the adver formance metrics which are absolute values. Merely as one tiser, calculated as the total cost for all clicks divided by example, CTR is a proportional performance metric since it the number of clicks; denotes the proportion between clicks (an ab solute value) and [0046] “Cost per impression”: the average cost of an impressions (another absolute value). As an alternative, a impression to the advertiser, calculated as the total cost proportional performance metric may be a proportion for all impressions divided by the number of impres between an absolute performance metric and another param sions; eter, such as time. As yet another alternative, a proportional [0047] “Click-through rate (CTR)”: the ratio between performance metric may be a certain mathematic manipula clicks and impressions of the ad entity, namelyithe tion of a proportion between two absolute performance met number of clicks divided by the number of impressions; rics; the “potential reach” is an example, since it is de?ned as [0048] “Conversions”: the number of times in which 1 minus the unique-impressions-to-reach ratio. users who clicked (or otherwise interacted with) the ad [0059] “Single advertiser budget”: A monetary amount entity have consecutively accepted an offer made by the which an advertiser is willing to spend on advertising, in a advertiser during a given time period (e. g. a day, an hour, certain advertising platform, on a group which includes mul etc.). For examples, users who purchased an advertised tiple ad entities. product, users who subscribed to an advertised service, [0060] “Health index”: A numerical value being a weighted users who downloaded a mobile application, or users average of multiple components. These components include who ?lled in their details in a lead generation form; one or more proportional performance metrics and one or US 2014/0379464 A1 Dec. 25, 2014 more absolute parameters associated with a pertinent ad [0066] Reference throughout this speci?cation to “one entity. An “ab solute” parameters is either a performance met embodiment” or “an embodiment” means that a particular rics or any other advertising-related parameter associated feature, structure, or characteristic described in connection with the pertinent ad entity. The weights in this weighted with the embodiment is included in at least one embodiment average may be assigned according to some business logic of the present invention. Thus, the appearances of the phrases provided by an advertiser, or proposed automatically by a “in one embodiment” or “in an embodiment” in various computerized system. In some embodiments, the common places throughout this speci?cation are not necessarily all characteristic of all the components which compose the referring to the same embodiment. Furthermore, the particu health index is that they each positively correlate to a mon lar features, structures, or characteristics may be combined in etary spending (and hence may each be referred to as being any suitable manner in one or more embodiments. monotonic with respect to spend). Namely, from a business [0067] As will be appreciated by one skilled in the art, perspective, a relatively lower value of such component will aspects of the present invention may be embodied as a system, justify a relatively lower money spend, and vice versa. For method, apparatus or computer program product. Accord example, for an ad entity exhibiting relatively low CTR, an ingly, aspects of the present invention may take the form of an advertiser would likely want to allocate a relatively low bud entirely hardware embodiment, an entirely software embodi get; in contrast, if the ad entity has a relatively high CTR, the ment (including ?rmware, resident software, micro-code, advertiser will likely wish to allocate a relatively high budget. etc.) or an embodiment combining software and hardware In other words, a high value of such component implies that aspects that may all generally be referred to herein as a “cir the ad entity is successful, and is worth investing more money cuit,” “module” or “system.” Furthermore, aspects of the in. present invention may take the form of a computer program product embodied in one or more computer readable medium DETAILED DESCRIPTION OF EMBODIMENTS (s) having computer readable program code embodied [0061] Disclosed herein is an advantageous budget distri thereon. bution in online advertising. Given a certain budget allocated [0068] Any combination of one or more computer readable by an advertiser for spending on multiple ad entities, an medium(s) may be utilized. The computer readable medium optimal distribution of the budget between the multiple ad may be a computer readable signal medium or a computer entities is computed. That optimal distribution may be com readable storage medium. A computer readable storage puted based on historical performance data of the multiple ad medium may be, for example, but not limited to, an elec entities, which is predictive of their future performance. tronic, magnetic, optical, electromagnetic, infrared, or semi [0062] The historical performance data may include mul conductor system, apparatus, or device, or any suitable com tiple performance metrics for each of the multiple ad entities. bination of the foregoing. More speci?c examples (a non Proportional performance metrics are those metrics which are exhaustive list) of the computer readable storage medium composed of a certain ratioia proportion between two (or would include the following: an electrical connection having more) values. Examples include click-through rate, conver one or more wires, a portable computer diskette, a hard disk, sion rate, return on investment, revenue per click, cost per a random access memory (RAM), a read-only memory impression, cost per click, revenue per impression, and poten (ROM), an erasable programmable read-only memory tial reach. (EPROM or Flash memory), an optical ?ber, a portable com pact disc read-only memory (CD-ROM), an optical storage [0063] Proportional performance metrics are usually an important factor considered by advertisers. Many advertisers device, a magnetic storage device, or any suitable combina require or aim for certain performance of their online adver tion of the foregoing. In the context of this document, a tising efforts, for which these proportional performance met computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in rics are often considered an excellent measure. At the same time, advertisers commonly allocate a set budget for their connection with an instruction execution system, apparatus, online advertising efforts, but fail to distribute this budget or device. between their ad entities in a way which maximizes those of [0069] A computer readable signal medium may include a the proportional performance metrics which are important to propagated data signal with computer readable program code the advertiser. embodied therein, for example, in baseband or as part of a [0064] Advantageously, in present embodiments, maximi carrier wave. Such a propagated signal may take any of a variety of forms, including, but not limited to, electro-mag zation of one or more of the multiple proportional perfor mance metrics may be achieved by way of the aforesaid netic, optical, or any suitable combination thereof. A com optimal distribution of the budget. The optimal distribution puter readable signal medium may be any computer readable may be either fully automatic, distributing the budget accord medium that is not a computer readable storage medium and ing to predetermined criteria, or be semi-automatic, by way of that can communicate, propagate, or transport a program for allowing an advertiser to manually de?ne an objective which use by or in connection with an instruction execution system, emphasizes one or more of the proportional performance apparatus, or device. metrics. [0070] Program code embodied on a computer readable [0065] In the following description, numerous speci?c medium may be transmitted using any appropriate medium, details are set forth to provide a thorough understanding of the including but not limited to wireless, wireline, optical ?ber embodiments. One skilled in the relevant art will recognize, cable, RF, etc., or any suitable combination of the foregoing. however, that the techniques described herein can be prac [0071] Computer program code for carrying out operations ticed without one or more of the speci?c details, or with other for aspects of the present invention may be written in any methods, components, materials, etc. In other instances, well combination of one or more programming languages, includ known structures, materials, or operations are not shown or ing an object oriented programming language such as Java, described in detail to avoid obscuring certain aspects. Smalltalk, C++ or the like and conventional procedural pro