CN101218597A - Methods and apparatus for generating, optimizing and managing granular advertising campaigns - Google Patents

Abstract

Systems for and methods of the present invention are directed to managing and optimizing ad campaigns. One method in accordance with the present invention comprises selecting a parent advertising campaign and generating a child advertising campaign, wherein the child advertising campaign automatically inherits selected advertising criteria from the parent advertising campaign. Another method in accordance with the present invention comprises determining performance metrics for multiple advertisements in an advertising campaign, selecting an advertisement from the multiple advertisements based on its performance metric, and running the selected advertisement. Performance metrics include, but are not limited to, return on ad spend, conversions, number of clicks on an advertisement, number of purchases, to name a few metrics.

Description

Methods and apparatus for generating, optimizing and managing granular advertising campaigns

related application

This application claims, pursuant to 35 U.S.C. §119(e), co-pending U.S. Provisional Patent Application No. 60/649,205, filed February 1, 2005, entitled "Method and Appargus for Generating, Optimizing and Managing Granular Advertising Campaigns" priority, which application is incorporated herein by reference.

technical field

The present invention relates to electronic commerce. More specifically, the present invention relates to electronic advertising campaigns conducted on the World Wide Web.

Background technique

Using keywords, web-based advertisements are better able to target advertisements to more likely customers. Web-based advertising also allows the merchant to track the effectiveness of the ad by quickly calculating the percentage of users who clicked through to the merchant's site and actually viewed the ad. The effectiveness of an advertisement may thus be calculated using marketing metrics such as return on advertising spend (ROAS), cost per click (CPC), and the like. Some services, such as Google ^™ 's AdWords(R), pay-per-click (PPC) services, allow merchants to specify which keywords will trigger their ads and how much they are willing to pay per click. Other services allow merchants to track return on investment (ROI). While services exist to generate campaigns and track ROI, there are no services that manage advertising campaigns by controlling multiple criteria for the campaigns.

Contents of the invention

The present invention relates to systems and methods for managing and optimizing advertising campaigns. In one aspect, a method of managing advertising campaigns includes selecting a parent advertising campaign and generating a child advertising campaign, wherein the child advertising campaign automatically continues advertising criteria selected from the parent advertising campaign. Preferably, the advertising criteria include at least one of keyword, creative and bid. In one embodiment, changing the value of an ad criterion in a parent campaign will automatically trigger a change in the value of that ad criterion in a child campaign. In another embodiment, the method further includes setting the parent advertising campaign to trigger when the first type of keyword matches, and setting the child advertising campaign to trigger when the second type of keyword matches.

In one embodiment, each of the first type keyword match and the second type keyword match is one of an exact match, a phrase match and a broad match. In yet another embodiment, the parent advertising campaign targets a first geographic location and the child advertising campaign targets a second geographic location different from the first geographic location. In another embodiment, the ratio between the value of an advertisement criterion in the parent campaign and the value of the advertisement criterion in the child campaign varies as a function of the relationship between the first geographic location and the second geographic location. It will be appreciated that the parent and shadow campaigns can each have any combination of advertising criteria, each independent of the other.

In a second aspect of the invention, a method of managing an advertising campaign includes determining a performance measure of a plurality of advertisements in the advertising campaign, and selecting an advertisement from the plurality of advertisements based on its performance measure. Each advertisement is formed by combining keywords and ideas, where each keyword and each idea has an associated rank. In one embodiment, the combination of keywords and ideas is determined by a back-off algorithm. Preferably, the selected advertisements are selected based on performance metrics associated with their match types. The match type is one of exact match, phrase match, and broad match. In another embodiment, the method further includes adjusting the bid for the keyword based on the match type. In another embodiment, the method further includes identifying new keywords to add to the advertising campaign based on the match type. In another embodiment, the method further includes determining from the plurality of keywords which keywords to run in selected ones of exact match, phrase match and broad match.

In one embodiment, each ad has a corresponding combination of creative, match type, landing page, and geographic targeting. Preferably, multiple advertisements run concurrently. Alternatively, multiple ads run sequentially.

In another embodiment, the method further includes generating a plurality of combinations of advertisement criteria for the plurality of advertisements. The performance measure for the selected advertisement corresponds to the conversion rate. In one embodiment, the method further includes adjusting the bid based on the performance metric corresponding to each of the plurality of advertising criteria. Preferably, the method further includes running a plurality of advertisements concurrently, each advertisement containing a keyword, and determining a performance measure for each advertisement. Alternatively, the method also includes sequentially running a plurality of advertisements, each advertisement containing a keyword, and determining a performance measure for each advertisement.

In another embodiment, the advertisement criteria include any one or more of keywords, channels, syndicates, creatives, match types, landing pages, geographic regions, days of the week, time of day, age, and gender.

In another embodiment, multiple advertisements are related in a tree structure with a parent node and corresponding child nodes. A parent node corresponds to an advertisement, and child nodes each correspond to a match type of the parent node advertisement. The method also includes deleting a child node from the tree when the performance metric corresponding to the parent node is below a predetermined threshold.

In another embodiment, the advertisement criteria correspond to a geographic target, and the method further includes determining a bid for each advertisement based on its geographic target and its corresponding performance metric.

In another embodiment, the method further includes determining a source of action for each of the plurality of advertisements, and removing advertisements that run from sources for which the ad performance metric is below a predetermined threshold. The source of the action is identified by an Internet Protocol address. Preferably, the method further comprises determining a referrer Uniform Resource Locator comprising an Internet Protocol address.

In another embodiment, the method further includes specifying a plurality of performance targets for the plurality of advertisements, and adjusting bids for the plurality of advertisements based on the plurality of performance targets. Performance goals include the maximum total spend for the entire campaign and the maximum spend for ads within the campaign.

In another embodiment, the method further includes determining a first keyword for the advertising campaign, automatically determining negative keywords for the first keyword, and only if the document triggering the advertising campaign contains the first keyword but does not contain the negative keyword Ads are only run from the campaign. Negative keywords are determined from at least one of search terms and conversion data.

The method also includes determining a chain of clicks to access the item via an advertisement in the advertising campaign, determining a value for the click in the chain of clicks, and assigning a performance measure to each click in the chain of clicks. Preferably, the allocation performance measure is based on any one or more of time to click, order of clicks in a chain of clicks, and number of ads clicked.

In a third aspect of the invention, a shadow campaign system includes means for generating a shadow campaign from a parent advertising campaign and means for populating the shadow campaign with advertising criteria selected from the parent advertising campaign. The shadow campaign module is configured such that its advertising criteria can be manually set. In another embodiment, the shadow activity is a selected one of a conditional shadow activity and an unconditional shadow activity.

In a fourth aspect of the present invention, a system for managing advertising campaigns includes: a first module for generating a plurality of advertisements, each advertisement containing one of a plurality of combinations of advertising criteria; A performance calculator for ad performance in advertising. Ad criteria include any two or more of geographic location, business site, and match type. Alternatively, ad criteria include any two or more of creative, landing page, and geotargeting criteria. Alternatively, the ad criteria include any two or more of keyword, channel, syndicate, day of the week, and time of day.

In another embodiment, the system further includes an execution module for executing the generated advertisements. Run modules configured to run generated ads concurrently. In some embodiments, the execution module invokes the system that displays the advertisement. In other embodiments, the runtime module itself displays advertisements. Alternatively, the run module is configured to run the generated advertisements sequentially.

In another embodiment, the system also includes means for removing advertisements that do not meet the threshold performance metric. In another embodiment, the system further includes means for determining a purchase item of an advertisement in an advertising campaign and means for determining a performance measure of a click in a chain of clicks leading to the purchase. The means for determining the performance measure is configured to analyze a click time in the train of hits, an order of clicks in the train of hits, and a number of clicks in the train of hits.

Description of drawings

Figure 1 is a high level block diagram showing several components of one embodiment of the present invention.

Figure 2 shows a parent activity and its corresponding shadow activity according to the present invention.

Figure 3 shows the steps of creating a parent campaign and its corresponding advertising campaign according to the present invention.

Figure 4 shows the steps for optimizing advertising performance in an advertising campaign according to the present invention.

5 is a table that maps keyword types to ranks that are paired sequentially with respect to keywords using an algorithm in accordance with the present invention.

Figure 6 shows an ad run using sequential paths in accordance with the present invention.

FIG. 7 is a table showing the number of paths generated based on specified criteria according to the present invention.

Figure 8 shows nodes in a tree showing advertisements in an advertising campaign according to the present invention.

Figure 9 is a high level diagram of the components that generate and run advertisements in accordance with the present invention.

Figure 10 shows the steps for creating an advertising campaign according to the present invention.

Figure 11 is an example of a weekly performance report generated according to the present invention.

Figure 12 is a table showing statistics for optimizing an advertising campaign in accordance with the present invention.

Detailed ways

Embodiments of the present invention are used to efficiently manage and optimize extremely fine-grained advertising campaigns, such as those created for search keywords and contextual advertising. As used herein, a granular advertising campaign is one in which an ad is targeted for a small number of impressions and the results are measurable in terms of impressions/clicks and back-end transactions. As used herein, an impression refers to any display of an ad. Preferably, the advertisement is interactive, so that a prospect can "click," visit, or otherwise interact with the advertisement, thereby triggering a report on the prospect's actions. This happens on the World Wide Web, and it can also happen on cellular phones, wireless devices, interactive TVs, interactive kiosks, and connected personal digital assistants (PDAs), just to name a few devices.

Embodiments of the present invention are used to efficiently create, optimize, or create and optimize large, complex and extremely granular advertising campaigns. According to one embodiment of the invention, activities are managed using shadow activities, also known as hierarchically related activities. According to this embodiment, the sub-campaign targets a subset of the total possible exposures. Since the results of each shadow activity can be tracked, rules can be generated to control how the shadow activity is set up relative to the parent activity.

Furthermore, according to the present invention, keywords and creatives are used to manage advertising campaigns. As used herein, Creative consists of the information used to display and manage an advertisement. This includes, for example, titles, descriptions, displays, click through rates, keywords and their associated bids. Keywords and creatives are matched, and metrics for each combination are generated. Subsequently, the best performing combination can be selected for display. Of course, the best performing combination can change based on, for example, changing objectives, changing product descriptions and prices, and changing business circumstances.

Additionally, according to the present invention, advertising campaigns are managed by tracking and analyzing match types. As used in this article, match type is the type of match that must occur before an ad can be displayed. In some embodiments, the match type is any of exact or exact match, phrase match, and generalized match, all of which are well known in the art. According to these embodiments, when an ad is displayed, it is determined by the targeting (eg, keyword or content category) and the match type used to match the ad targeting to the displayed category. If the match is exact, the ad will only be shown if the displayed category exactly matches the ad's goals. If the match is a broad match, the rule is used to determine what ad to show on which impression.

According to other embodiments of the invention, negative keywords are automatically generated. For example, for "broad" and "phrase" match types, negative keywords are automatically generated based on an analysis of search term information in conversion data, which is data that indicates when the display of an ad prompted a user to visit it.

According to other embodiments of the invention, advertising campaigns are managed by executing sequential paths. In these embodiments, trials are run sequentially when the ad network does not support separate and concurrent keyword instances within a campaign, and comparison between trial ideas is not possible.

Other embodiments of the invention relate to optimizing advertising campaigns. In one of these embodiments, keyword paths (eg, "local paths") are optimized. In these embodiments, each keyword, creative, channel, match type, syndication, and several other factors that control where and how ads are displayed create paths or different combinations of variables. Optimization is determined over all possible combinations. Determining how to strengthen or reduce the strength of a particular path depends on several factors.

According to an embodiment of the invention, an advertising campaign is optimized by performing a global optimization. In these embodiments, performance goals (eg, return on advertising spend or "ROAS" goals) are each set at the product/category, campaign, ad group, and keyword/creative levels throughout the campaign hierarchy. When a goal is set at a certain level, optimizations for each descending level are automatically calculated to achieve the goal.

In other embodiments, campaign advertisements are optimized by removing keyword paths. In these embodiments, the entire campaign structure is selectively deleted based on the analysis of performance metrics at each level of the hierarchy in order to keep the total number of keywords within technical and practical constraints. The total number of keywords can be set to any maximum value specified per active user.

In other embodiments, campaign advertisements are optimized by analyzing geographic data. National events are divided into metropolitan areas. The performance of each keyword path at the metropolitan level is measured and subsequent bids are derived accordingly.

In other embodiments, campaign advertising is optimized by analyzing the business site. The business performance of each publisher site is tracked and if the site fails to meet a certain metric or falls within some specified number/section of the lowest performing sites, the site is removed from the syndicate.

In other embodiments, campaign advertisements are optimized by optimizing multiple objectives. In these embodiments, a number of goal constraints are specified. These constraints include per-order cost constraints and monthly budgets.

In other embodiments, campaign advertising is optimized by optimizing the purchase funnel. Users typically view and click on multiple advertisements before making a purchase. In these embodiments, the contribution of earlier clicks relative to the final purchase is calculated and the value of each purchase is attributed to its earlier ("head") item.

Embodiments of the invention thus relate to campaign management, optimization and reporting. Embodiments contain "glue logic" to interface with many ad networks and tracking systems and simplify the view of ad campaigns to administrators, while building complex campaign structures internally to fully utilize all available targeting mechanisms provided by ad networks .

For example, Figure 1 shows a high-level diagram of a system 100 in accordance with the present invention. System 100 includes campaign management module 101 and optimizer module 103 , both connected to data warehouse 105 containing campaign structure 107 , click data 109 and conversion data 111 . Campaign structure 107 is used to define and manage advertising campaigns; click data 109 is used to track click-through data and the like; and conversion data 111 is used to track conversions. To simplify the discussion below, the term "management" is used to refer to a module capable of managing or optimizing, or both, a campaign advertisement.

event management

shadow activity

A shadow activity is a way to form a new dynamic copy of an activity and change a small part of the new activity. They are especially useful for geographic targeting and syndicate level and match type differentiation.

A shadow activity works as follows: First, the parent activity is identified, and a shadow (eg, child) activity is generated and assigned a name. Next, copy all keywords and ideas from the parent activity in the shadow activity. Bids from the parent campaign are multiplied by a prescribed ratio to generate equivalent bids in the shadow campaign. The match type in the shadow activity is optionally set to a value equal to or different from the value in the parent activity. For example, a parent activity could be set to phrase match for all keywords, while a shadow activity could be set to broad match.

Later, any changes made in the parent activity (for example, adding new keywords or modifying ideas) will be reflected in the shadow activity, unless the corresponding entity has changed in the shadow activity. For example, if an idea was first modified in the shadow activity, changing that idea in the parent activity will not result in a change in the shadow activity.

Preferably, the bid ratio is considered dynamic. For example, if the shadow is set to have 75% of the parent's bid, when the parent's bid updates, the shadow's bid updates to 75% of it. On the other hand, if a keyword's bid is manually set to a certain value, changing the bid in the parent will not update the bid in the shadow. As shown in Figure 2 campaign series 200, an example of an application of this aspect would be to create a Canada-targeted campaign (shadow campaign) from a US campaign (parent campaign). Campaign series 200 shows US campaign (e.g., parent campaign) 201 with (1) title "US campaign" 203, (2) first keyword block 205 titled "Keyword 1" 205A with Creative 1 205B and bid 205C, And (3) a second keyword block 207 titled "Keyword 2" 207A with Idea 2 207B and Bid 207C. US campaign 201 has Canada shadow campaign 210 with: (1) title "Canadian campaign" 213; (2) first key titled "Keyword 1" 215A copied from US campaign 205 first key block Block 215 with creative C1 215B and a bid 215C corresponding to 75% of the bid value in the first keyword block for US campaign 205C; and (3) titled "Keyword 2" copied from US campaign 207 second keyword block A second keyword block 217 of 217A, with a creative C 2217B and a bid 217C corresponding to 75% of the bid value in the second keyword block for the US campaign 207C.

Bids in Canadian campaigns defaulted to 75% of U.S. bids, except for creative changes. One advantage of this option is that keywords added to the US campaign are automatically added to the Canadian campaign.

It will be appreciated that shadow structures can be applied at any level in the activity hierarchy. Some embodiments of the invention support shadow products/categories, campaigns and ad groups.

FIG. 3 shows a step 300 of generating a shadow activity from a parent activity according to an embodiment of the present invention. First, in step 301, select the parent activity, and then in step 303, select the inherited property to be copied to the shadow activity. In step 303, factors such as bidding multiples are also determined. In step 305, shadow activity is then determined.

Figure 4 shows a sequence of steps 320 for managing an advertising campaign in accordance with the present invention. In a first step 321 performance metrics are collected for a plurality of advertisements in an advertising campaign. Performance metrics include, but are not limited to, return on advertising spend (RAOS), cost per action, number of actions, return on investment (ROI), revenue, or any other metric used to measure advertising or advertising campaign performance. Next, in step 323, the advertising campaign is managed, such as by creating advertisements to run or optimizing advertisements. Finally, in step 325, the selected one or more advertisements are run.

In another embodiment, the parent activity is a superset of the shadow activity. This parent and shadow relationship has advantages, especially for key path removal, described in more detail below. For example, a parent activity is a US-wide activity and has New York and Los Angeles shadow activities. If the volume of a given keyword is low, causing the shadow campaign to be removed, the parent campaign will still be able to cover the New York and Los Angeles metro areas with a nationwide campaign.

Keyword and creative rating

According to other embodiments, ad text affects click through rate (CTR) and conversion rate/ROI (CVR). For example, the ad "Free film processing, free shipping" will likely have a higher CTR and lower CVR than the ad "24 cents per print, good archive paper". The first ad text or creative is called "aggressive" because it mentions the word "free," while the second ad text or creative is called "conservative" because it mentions price.

As used herein, "creative" refers to the information used to create and track an ad, such as the ad title, description, display, click-through URL, keywords, and bid.

Based on the keyword, it can be determined to run an aggressive creative on it. Aggressive creatives will maximize business, but also maximize advertising spend. Embodiments of the present invention allow campaign managers to specify a rating for each keyword and each creative. Therefore, if a keyword is rated positive and a positively rated idea is available, the keyword and idea will be paired together.

When no exact rank match exists, a backoff algorithm as shown in table 400 of FIG. 5 is used. Table 400 contains rows 405, 407, 409, and 411, where entries in column 401 indicate the level type of the key (KW), and entries in corresponding column 403 indicate the fallback algorithm. For example, row 405 contains entries indicating that when no keywords are specified (column 401), all ideas are used by using an algorithm (column 403). Row 407 contains entries indicating that when the keyword is aggressive (column 401), the order of pairing by using the algorithm is as follows: it is first paired with the aggressive idea; if the aggressive idea is not present, it is paired with the unspecified idea; If the unspecified idea does not exist, it is paired with the neutral idea; if the neutral idea does not exist, it is paired with the conservative idea (column 403). Row 409 contains entries indicating that when the keyword is neutral (column 401), the pairing order by using the algorithm is as follows: it is paired with the neutral idea first; if the neutral idea does not exist, it is paired with the unspecified idea; If the unspecified idea does not exist, it is paired with the conservative idea; if the conservative idea does not exist, it is paired with the aggressive idea (column 403). Row 411 contains entries indicating that when the keyword is reserved (column 401), the order of pairing by using the algorithm is as follows: it is first paired with the conservative idea; if the conservative idea does not exist, it is paired with the unspecified idea; If the unspecified idea does not exist, it is paired with the neutral idea; if the neutral idea does not exist, it is paired with the positive idea (column 403).

Match Type Analysis

Some embodiments of the invention exploit match types using two techniques for optimizing advertising campaigns. In these embodiments, particularly those running on Google, the same keyword was run three ways (ie, once with each match type) across multiple campaigns. Keywords are run only once in broad match. The tracking system identifies actual search terms entered at runtime (eg, using referral link information from hypertext transfer protocol, or "HTTP"), and tracks return on investment (ROI) for each actual search term. The accumulated search terms (and their results) are then grouped according to whether they are an exact match, a phrase match, or a broad match of the keyword in the campaign. Ad spend, revenue, and conversion rates can then be identified for each match type for each keyword. The match type information can then be used to adjust bids and also identify new keywords to be added to the campaign, or which existing keywords should be explicitly run in exact or phrase match mode.

Negative keywords are automatically generated

When "broad" or "phrase" keyword matching is used with a channel, it is often necessary to use negative keywords to more appropriately contextualize the ad slot. For example, a seller of women's shoes may bid on the keyword "shoes." However, a "broad" or "phrase" match could place the ad with key terms such as "brake pads" or "horseshoes." Specifying 'brakes' or 'horses' as negative keywords ensures that the seller's ad does not appear in such undesirable contexts.

Negative keywords are often unpredictable. Negative keyword auto-generation evaluates actual search terms derived from traffic and conversion data, and based on its analysis will determine which search terms to include as negative keywords.

sequential path

For some optimization processes, it is best to combine the same keyword with multiple concurrent experiments for creative, match type, landing page (the first page a customer encounters when visiting a site, which can be different from the site's home page), or geographic Targeting criteria run together. The optimization of the campaign can then be fine-tuned based on the performance comparison of each trial. However, when a channel does not support concurrent experiments with the same keyword, these experiments will run sequentially, rather than concurrently. Manage the timing and sequence of trials to produce similar results.

For example, if three ideas are related to a particular keyword, they can run one idea at a time for three consecutive months. The results can then be analyzed to determine which idea to select. Preferably, the other two ideas run automatically on a regular basis (eg, one week per quarter) as user behavior changes over time.

Figure 6 shows a sequence 500 of paths 501-505 run sequentially in accordance with the present invention. Path 501 run in January uses creative CR1 and a target return on ad spend (ROAS) of 100% for keyword KW1 ^* . Next in order, run path 502 in February for keyword KW1 ^* using creative CR2 and target ROAS of 300%. Paths 503-505 have parameters with values specified as such.

optimization

Keyword Path (Local) Optimization

Keyword path optimization refers to the consideration of multiple factors in determining bids, keywords, and creative choices. All sensitive portfolios are listed, and the return on investment for each portfolio is measured, and bids are priced accordingly. Several factors include, but are not limited to, channels, syndicates, keywords, and channels, each of which will be discussed in turn.

Channels: Typically, each channel has a different bid price for the same keyword. For example, the bid price for a keyword on Google will be different than the bid price for the same keyword on Overture.

Syndicate: Specifically for Google, it bids and measures performance for Google.com, search partner, and content partner businesses separately. Yahoo (Overture) allows separation of search and content site businesses.

Keywords: Each keyword bids and performs differently. Even misspellings and plurals can have huge performance differences.

Creative: Multiple ads should be run whenever possible, and each ad should be treated as a separate portfolio.

Match Type Optimization: Match Type is a construct used by the Search Network to increase the distribution of campaigns without exhaustively specifying all matching keywords. One embodiment of the present invention runs each keyword in each match type and calculates the appropriate cost-per-click ("CPC") bid for each variant.

Syndicate Optimization: An advertising network is literally a network of hundreds or thousands of websites. In an ideal world, the business performance of individual websites should be measured and bid on. This is not always possible, but generalized groupings can be achieved, such as search sites and content sites.

Day/Time Optimization: This optimization measures campaign performance based on the most recent Monday, Tuesday, etc. and adjusts bids accordingly.

Landing page optimization: For customers with multiple possible landing pages, all pages are preferentially used and measured.

Keyword path optimization can be truly local. According to one embodiment, only one path is viewed and optimized independently of any other paths. According to other embodiments, keyword path optimization also extends to related paths, such as when optimizing all paths derived from a single keyword.

FIG. 7 shows a table 600 showing path numbers for typical activities according to one example. Table 600 contains rows 601-610 and columns 651-655. Each column 651-655 is labeled to indicate the type of entry in the column. The entries in row 601 are headings describing what the value in a particular column indicates. Thus, for example, the column labeled "Criteria" (651) contains an entry for each criterion describing a portion of the campaign, such as "Keywords" (row 602) and "Channel" (row 603). The column labeled "Selection" (652) indicates the number of choices for that particular criterion. So, for example, row 602 is for the "Keywords" criterion (column 651), which in this example has 2000 choices (column 652), with the component "Singles" ( column 653), and thus has 2000 associated paths (column 655). Similarly, row 603 is for the "Channel" criterion (column 651), which in this example has 6 choices (column 652), with the component "Google" ( column 653), and thus have 12000 associated paths (column 655) determined by multiplying the 2000 choices of the "keyword" criterion by the 6 choices of the channel criterion. The remaining examples are illustrated in a similar fashion.

global optimization

Global optimization refers to setting performance goals at a higher level and adjusting the goals of deeper active entities in such a way as to best achieve the higher level goals. For example, the entire account is to be optimized for 200% ROAS. (This is also known as a portfolio-level goal.) In a simple case, the global optimizer would want each product/campaign/adgroup to perform at the 200% level.

However, since some campaigns (for example, those containing branded items) can consistently perform better than 200%, others may only need to achieve 180% ROAS for the portfolio to achieve its goals.

Similarly, while a keyword-level goal of 200% can be set, a keyword may consistently perform at 150% for whatever reason. So, to hit the actual target of 200%, maybe 250% of the indicator target may be necessary. The global optimizer therefore needs to tune the metric target to achieve a realistic goal.

Keyword path removal

The process of optimization requires duplicating keywords across channel, campaign, and syndicate levels, each with variations in creative, match type, and bid. Taking the example of a flat text file with a separate line for each variation of each parameter for each keyword, the number of line items can quickly exceed the practical and technical limits that an ad network will allow.

For example, Google has a 100,000 line limit on activity definitions. As detailed in the table 600 of Figure 7, it can be seen that expanding the paths for match type, metro area, day of week and time of day would push the campaign beyond this limit. Therefore, it is necessary to limit the total number of paths.

One solution is that keyword path removal analyzes traffic and conversion data, and expands or removes the path tree based on whether there are enough conversions to still make sense to split the buckets and whether there is any benefit to be gained from that split. Here, the term "route tree" refers to a tree structure in which nodes refer to advertisements, and some nodes ("child nodes") are created by adding advertisement criteria to a "parent node".

As an example of a rule of thumb, if a node in the tree only had 10 conversions in the last month, it may be determined that it should be split further. As another example, if the analysis shows similar conversion rates for all three match types for this keyword, it is decided not to perform a match type split (ie, in a campaign that has the same keyword in exact, phrase, and broad match forms). In such cases, the keyword operates in broad match mode.

Figure 8 shows a tree 700 used to describe one embodiment of the present invention. Tree 700 contains node 701 and node 703 . Node 701 is for the keyword KW ^* which has the idea "idea 1" of 9 orders. The path terminates at node 701 because the number of orders (9) for this advertising campaign is below a predetermined threshold.

In contrast, node 703 is for keyword KW5 ^* with creative "creative 2" and has 100 orders. Therefore, the path does not terminate there; instead, node 703 has three child nodes 705, 707, 709 with parameters as shown in FIG. 7 .

geographic analysis

Similar to the match type analysis feature, geographic analysis can be performed on the results. This is also done in explicit and implied mode. In explicit mode, individual campaigns are created and targeted to specific countries or metro areas based on what the ad network provides for targeting capabilities. In implicit mode, the Internet Protocol (IP) address of the referenced link or the IP address of the user can be used to reverse locate the source. Origins can be located in reverse using any number of means, including but not limited to reverse IP addressing.

Business Site Analysis

By checking the IP address of incoming traffic, a tracking system built into the advertiser's website can identify the site from which the click came. Conversion performance by source site can then be analyzed.

This data was then used in two ways: (1) low performing sites were identified and excluded from the list of sites used. (2) Since some ad networks allow different CPC bidding rules by site, the CPC is set to a value commensurate with the conversion behavior of the site.

multiple optimization goals

A client may specify multiple performance goals or constraints. These goals or constraints are generally based on budget and ROI. For example, "Optimize bids down so that we spend no more than $30,000 total this month, or $10 per order - whichever is lower". Sometimes, users want to "whichever is higher" instead.

If a customer is looking to buy a new TV, he or she might first search for the item "color tv," then "plasma tv," then "sony tv," then "sony kvm4542," and then click on each Ads on results pages.

Optimization systems typically attribute the entire purchase value to the last-clicked ad. According to the present invention, the value among the clicked advertisements is apportioned based on several factors including: how recently the advertisements were clicked; the order of clicks; and the number of clicked advertisements.

The purpose is to appropriately increase the value of the early "head" item and appropriately decrease the value of the later "tail" item.

Figure 9 is a high level diagram of a system 720 in accordance with the present invention. System 720 includes a management module 721 connected to an execution module 723 . Here, management is broadly defined to include managing and optimizing advertising campaigns to fit existing applications.

In operation, embodiments of the present invention utilize hierarchy and inheritance. Elements such as bids and goals can be set at a high level and then inherited. Alternatively, any inherited values can be explicitly overridden by setting values at lower levels in the path tree as needed.

In one embodiment, the data hierarchy is in descending order:

• Accounts: Many accounts can be processed as needed.

• Channels: Each ad network can be viewed individually, or the entire view across all channels can be added together.

• Category/Product: Using this level, activities can be subdivided into arbitrary units representing meaningful divisions for the enterprise. These cells include product listings.

• Campaign/Targeting: At this level, geographic targeting (both country and metropolitan) is performed, as well as shadow campaigns.

· Ad Group: A public list of related keywords displayed as a collection of public creatives.

· Keywords and ideas.

Figure 10 shows a sequence of steps 750 for creating an advertising campaign in accordance with the present invention. First, in step 751, the user creates a new account. In this step, for example, when search and content targeting campaigns are created, they will default to a ratio based on the bids in the parent campaign.

In step 753, products, categories, or both are selected to add to an advertising campaign, and in step 755, are used to create an advertising campaign. In step 753, the product and category have been given meaningful names, and the maximum bidding cost per click has been selected. A campaign is created in step 755, such as by using Google's AdWords. In one embodiment, the first letter of the campaign name is the syndication level (e.g., G, S, C in Google, Search, and Content) and the second letter is the match type (e.g., exact, phrase, and broad X, P, B in). The remaining letters are geotargeting information. So, for example, "SBUS+CA" refers to a broad search targeting the United States and Canada. Metropolitan-level activities are conditional shadow activities. The SBUS+CA activity is the parent activity, which can be manually modified.

Next, in the step, create an ad group for the campaign. Ad groups can be created on all channels. Ad groups can have multiple creatives, which can run concurrently or sequentially. Embodiments of the present invention select and run the highest performing creative for each keyword.

Many campaign systems have limitations on the size of the campaign. Embodiments of the present invention are used to limit campaign size by, for example, using conditional shadow campaigns (CSC). A CSC is a shadow activity that is created when a predetermined number of conversions from its parent activity have taken place. Such selective shadow activity creation is similar to deletion above.

Embodiments of the present invention also include a reporting feature that can generate periodic reports to show ad performance for each week, month, quarter, or year relative to a calendar, or day-of-week reports to show ad performance by day of the week. These reports allow ad owners to determine which ad campaigns are worth keeping and which should be replaced. Some of the statistics included in the report are as follows:

• Return on Ad Spend (ROAS), which is revenue divided by media spend, ie, how many dollars in revenue are tracked for each dollar spent on advertising.

• Cost per action, which is the amount the ad spends per tracked action. The actions are different for each client.

• Number of actions tracked during the reporting period.

· Average Single Order, which is the total amount earned divided by the number of orders.

· Cost per order, which is the total ad spend divided by the number of orders.

• The total number of orders for the reporting period.

· Order rate, also known as conversion rate, which is the total number of orders divided by the total number of clicks.

· The total number of active hits for the reporting period.

• The total number of proceeds or amounts tracked during the reporting period.

• Cost or total advertising spend for the period.

Figure 11 shows a table 800 for a weekly report according to one embodiment of the invention. Table 800 contains row 801 detailing a weekly report of the performance of all products in the campaign, row 803 detailing the performance of a single product "Avatar", and row 805 showing similar performance metrics for the channel GoogleAdwords. As used herein, channels are advertising networks such as Google, Overture, and Enhance. Row 801 has entries 801A-801J, and row 803 has entries 803A-803J, defined by headings above each column. For example, row 801 shows that all products together have a ROAS of 150% (entry 801A), a cost per action of 38 points (entry 801B), have 2679 actions (entry 801C), a 26.61% action rate (entry 801D), each Orders averaged $10 (entry 801E), resulted in $6.69 per order (entry 801F), generated 152 orders (entry 801G), had a 1.51% order rate (entry 801H), had 10066 clicks (entry 801I), and generated revenue of $1520.00 (Item 801J). Entries 803A-J show corresponding values for a single product "Avatar".

According to the present invention, target value and type can be set. Target types include: (1) ROAS, where the target value is a percentage; (2) the position or average position of an ad relative to competitor ads; (3) cost per order (CPO), where the target value is dollars and cents; (4) Cost per action; (5) Cost per order and action (CPOA), where the system first looks at the order and if there are insufficient orders, checks if there have been enough actions during the booking period. If enough actions are already available, the system converts actions into orders by using the overall action-to-order ratio for the product and category, and optimizes the resulting number of orders; and (6) closes, any optimized are closed.

Goal setting according to the present invention represents marginal goals on a per-keyword-path basis. For example, if the cost per order is set to $10, the optimizer should spend no more than $10 for the most expensive order. Overall campaign performance can deviate from this target, for example, certain paths (eg, "brand" projects) perform so high that even at the top positions, they exceed the target based on their return on investment (ROI). In other words, spending more on an order does not result in more orders. Therefore, ROI is higher than the target, so that while ROAS increases, CPO or CPOA decreases. As another example, a large set of keywords may not generate orders, but each do not generate enough business volume to allow them to be bid down or disabled.

Figure 12 shows a table 900 showing metrics for keywords used in campaigns managed in accordance with the present invention to illustrate how campaigns are optimized. This example assumes a target ROAS of 100. Table 900 contains rows 901, 903 and 905, each showing statistics for keywords in entries 901A, 903A and 905A, respectively. Table 900 is divided into statistics for a 7-day period 950 and statistics for a 30-day period 960, as described below.

Row 901 shows that the ad for "keyword 1" (entry 901A) generated 32 orders (entry 901B) within a 7-day period with a corresponding ROAS of 57.91 (entry 901C). Statistics from the 7-day portion of the chart are used because the number of orders in a 7-day period exceeds a predetermined threshold. Since the ROAS (57.91) is much smaller than the assumed ROAS of 100, the bid for this keyword is reduced by, say, 40%. Row 903 shows that the ad for "keyword 13" (entry 903A) generated 4 orders (entry 901B) within a 7 day period, a value below a predetermined threshold, so statistics from the 30 day portion 960 of the table are used to allow for a large enough Sampling provides meaningful data. The listed ROAS (entry 903E) for 30 days is 62.8, which is also well below the assumed ROAS of 100, so again the bid for this keyword is reduced by, say, 40%. Row 905 shows that the ad for "keyword 15" (entry 905A) generated 23 orders (entry 905B) over a 7 day period, a value above a predetermined threshold. The corresponding ROAS (entry 905C) is as high as 5205.41, but the weekly increment has been limited to a maximum of 100%. However, since the ad's rank is already 1 (entry 905D), no increase is required.

In operation, the user creates an advertising campaign. Multiple ads are created from this campaign. The performance metric associated with each ad is measured, and the ad with the highest performance metric is selected and run. In this way, the owner of an advertising campaign ensures that only the best performing ads are run, thereby ensuring that the owner maximizes profits. In other embodiments, poorly performing ads are not run, thereby allowing the owner to reduce any loss (cost of running the ad: profit). Embodiments of the present invention allow the owner to easily monitor, manage and create multiple advertisements to run against the advertisement campaign.

Embodiments of the invention can run on a variety of platforms, including but not limited to personal computers, cellular telephones, interactive televisions, interactive kiosks, and personal digital assistants.

It will also be appreciated that while the above discussion describes individual functions, some embodiments of the invention are capable of performing any combination of functions to manage advertising campaigns. For example, some embodiments of the invention are capable of performing any combination of generating shadow activities, selecting advertisements based on performance metrics, deleting child nodes in a tree structure, running multiple advertisements concurrently or sequentially, and collecting performance metrics, among others.

It will be readily understood by those skilled in the art that various other modifications may be made to the embodiments without departing from the spirit and scope of the present invention as defined in the appended claims.

Claims (47)

1. A method of managing an advertising campaign comprising: a. Select a parent campaign; and b. Generate a child advertising campaign, wherein the child advertising campaign automatically inherits the advertising criteria selected from the parent advertising campaign. 2. The method of claim 1, wherein the advertising criteria includes at least one of keywords, creatives, and bids. 3. The method of claim 1, wherein changing the value of an advertising criterion in the parent campaign automatically triggers a change in the value of the advertising criterion in the child campaign. 4. The method of claim 1, further comprising setting the parent advertising campaign to trigger on a first type keyword match and setting the child advertising campaign to trigger on a second type keyword match. 5. The method of claim 4, wherein each of the first type of keyword match and the second type of keyword match is one of an exact match, a phrase match, and a broad match. 6. The method of claim 1, wherein the parent advertising campaign targets a first geographic location and the child advertising campaign targets a second geographic location different from the first geographic location. 7. The method of claim 6, wherein the ratio between the value of the advertising criterion in the parent campaign and the value of the advertising criterion in the child campaign varies with the difference between the first geographic location and the second geographic location. The relationship between them changes. 8. A method of managing advertising campaigns comprising: a. determining performance metrics for a plurality of advertisements in said advertising campaign; and b. Selecting an advertisement from the plurality of advertisements based on its performance measure. 9. The method of claim 8, wherein each advertisement is formed by combining keywords and creatives, wherein each keyword and each creative has an associated rating. 10. The method of claim 9, wherein combinations of keywords and ideas are determined by a back-off algorithm. 11. The method of claim 8, wherein the selected advertisements are selected based on performance metrics associated with their match types. 12. The method of claim 11, wherein the match type is one of an exact match, a phrase match, and a broad match. 13. The method of claim 11, further comprising adjusting bids for keywords based on the match type. 14. The method of claim 11, further comprising identifying new keywords to add to the advertising campaign based on the match type. 15. The method of claim 11, further comprising determining, from a plurality of keywords, which keywords to run in selected ones of exact match, phrase match, and broad match. 16. The method of claim 8, wherein each advertisement has a corresponding combination of creative, match type, landing page, and geographic targeting. 17. The method of claim 16, wherein the plurality of advertisements run concurrently. 18. The method of claim 17, wherein the plurality of advertisements run sequentially. 19. The method of claim 8, further comprising generating a plurality of combinations of advertisement criteria for a plurality of advertisements. 20. The method of claim 19, wherein the performance measure of the selected advertisement corresponds to a conversion rate. 21. The method of claim 19, further comprising adjusting bids based on performance metrics corresponding to each of the plurality of advertising criteria. 22. The method of claim 21, further comprising running the plurality of advertisements concurrently, each advertisement containing a keyword, and determining a performance metric for each advertisement. 23. The method of claim 21, further comprising sequentially running the plurality of advertisements, each advertisement containing a keyword, and determining a performance metric for each advertisement. 24. The method of claim 19, wherein the advertising criteria include any of keyword, channel, syndicate, creative, match type, landing page, geographic region, day of the week, time of day, age, and gender or multiple. 25. The method of claim 21 , wherein the plurality of advertisements are related in a tree structure having a parent node and corresponding child nodes, wherein the parent node corresponds to an advertisement, and the child nodes each correspond to the The match type of the ad for a parent node, the method further comprising deleting a child node from the tree when a performance metric corresponding to the parent node is below a predetermined threshold. 26. The method of claim 19, wherein the advertisement criteria correspond to a geographic target, the method further comprising determining a bid for each of the advertisements based on its geographic target and its corresponding performance metric. 27. The method of claim 8, further comprising: a. determining a source of action for each of the plurality of advertisements; and b. removing advertisements running on sources for which performance metrics for said advertisements are below a predetermined threshold. 28. The method of claim 27, wherein the source of action is identified by an Internet Protocol address. 29. The method of claim 28, further comprising determining a reference link URL comprising an Internet Protocol address. 30. The method of claim 8, further comprising: a. specifying a plurality of performance targets for the plurality of advertisements; and b. Adjusting bids for the plurality of advertisements based on the plurality of performance goals. 31. The method of claim 30, wherein the performance goals include a maximum total cost for an entire advertising campaign and a maximum cost for an advertisement within the advertising campaign. 32. The method of claim 8, further comprising: a. determining a first keyword for said advertising campaign; b. Automatically determine negative keywords for the first keyword; and c. Running an ad from the ad campaign only if the document that triggered the ad campaign contained the first keyword but did not contain the negative keyword. 33. The method of claim 32, wherein the negative keywords are determined from at least one of search terms and conversion data. 34. The method of claim 8, further comprising: a. Identify the sequence of clicks through which the item was accessed through an ad in said advertising campaign; b. determining a value for said click in said series of clicks; and c. assigning a performance measure to each of said clicks in said chain of clicks. 35. The method of claim 34, wherein assigning performance metrics is based on any one or more of time to click, order of clicks in the chain of clicks, and number of ads clicked. 36. A shadow activity system comprising: a. A component for generating a shadow campaign from a parent campaign; and b. A component for populating said shadow campaign with advertising criteria selected from said parent advertising campaign. 37. The system of claim 36, wherein the shadow activity is a selected one of a conditional shadow activity and an unconditional shadow activity. 38. A system for managing advertising campaigns comprising: a. a first module for generating a plurality of advertisements, each advertisement comprising a combination of advertising criteria from among the plurality of combinations of advertising criteria; and b. A performance calculator for calculating advertisement performance from the plurality of advertisements. 39. The system of claim 38, wherein the advertisement criteria include any two or more of geographic location, business site, and match type. 40. The system of claim 38, wherein the advertisement criteria include any two or more of creative, landing page, and geolocation criteria. 41. The system of claim 38, wherein the advertisement criteria include any two or more of keywords, channel, syndicate, day of week, time of day, age, and gender. 42. The system of claim 38, further comprising an execution module for executing the generated plurality of advertisements. 43. The system of claim 42, wherein the execution module is configured to run the generated advertisements concurrently. 44. The system of claim 42, wherein the execution module is configured to execute the generated advertisements sequentially. 45. The system of claim 38, further comprising means for deleting advertisements that do not meet a threshold performance metric. 46. The system of claim 38, further comprising: a. a means for determining the purchase of an item from an advertisement in said advertising campaign; and b. means for determining a performance measure of clicks in a chain of clicks leading to said purchase. 47. The system of claim 46, wherein the means for determining a performance metric is configured to analyze the timing of the clicks in the chain of hits, the order of clicks in the chain of hits, and the time of clicks in the chain of hits. The number of clicks.

Images