Orthogonal corpus index for ad buying and search engine optimization
Systems and methods for applications of orthogonal corpus indexing (OCI), such as selecting ad words for purchase and improving visibility of web pages in search engines, are described. In one aspect, the systems and methods described herein employ OCI to recommend to an advertiser ad words for purchase. Advertisers pay search engines for placement of their advertising along side results in the search results page, when a given word or phrase appears in a user's search query. The described systems and methods enable automated selection of related and discriminating terms, identifying keywords that increase the ratio of ads clicked-through to money spent on keyword buying. In another aspect, the systems and methods described herein employ OCI to generate content for web pages. OCI may be used to determine content that when added to a web page improves the rank of that page in a search engine.
1. A system for selecting ad words for purchase, comprising:
an information database; and
a processor configured for:
processing the information database using orthogonal corpus indexing and a seed topic to derive a plurality of candidate ad words;
receiving an estimated upper and an estimated lower cost per click (CPC) value for each of the plurality of candidate ad words;
computing estimated upper and estimated lower marketing break even (MBE) values for each candidate ad word of the plurality of candidate ad words based on corresponding estimated upper and lower CPC values, wherein each estimated MBE value of a corresponding candidate ad word represents an estimated cost of an advertisement relating to the corresponding candidate ad word;
computing an average MBE value for each of the plurality of candidate ad words based on corresponding estimated upper and estimated lower MBE values;
selecting one or more ad words from the plurality of candidate ad words as a function of average MBE values associated with the selected one or more ad words;
receiving a measured upper and a measured lower CPC value for the one or more selected ad words based on an advertisement deployed relating to the one or more selected ad words, wherein the measured upper and the measured lower CPC values represent actual costs for one click on an advertisement related to the one or more selected ad words;
computing a measured average MBE value for the one or more selected ad words based on the measured upper and measured lower CPC values, wherein the measured MBE values for the one or more selected ad words represent actual cost of the advertisement relating to the one or more selected ad words; and
removing an ad word from the one or more selected ad words that has a measured average MBE value higher than a threshold.
2. The system of claim 1 , wherein the processor is configured for:
computing a global MBE average value based on the average MBE values for the plurality of candidate ad words, wherein the global MBE average value represents an average value computed across the plurality of candidate ad words; and
selecting a set of ad words from the plurality of candidate ad words that have their associated average MBE values below the global MBE value.
3. The system of claim 1 , wherein the selected one or more ad words include at least one of a positive ad word and a negative ad word, an advertisement based on the selected one or more ad words being displayed in response to the positive ad word, and prevented from being displayed in response to a negative ad word.
4. The system of claim 1 , wherein CPC estimates for each of the plurality of candidate ad words are provided by a traffic estimator based on number of clicks to an advertisement shown in response to a search query in a search engine having at least one of the plurality of candidate ad words.
5. The system of claim 1 , wherein the estimated upper and the estimated lower MBE values for each of the plurality of candidate ad words are computed based on one or more metrics from the group consisting of cost per click, clicks per unit time, conversion rate per click, conversion value, and cost per impression.
6. The system of claim 1 , wherein the plurality of candidate ad words includes keywords related to the seed topic and extracted from the information database.
7. A server for selecting ad words for purchase comprising:
a processor configured for:
processing information from an information database using orthogonal corpus indexing and a seed topic to derive a plurality of candidate ad words;
receiving an estimated upper and an estimated lower cost per click (CPC) value for each of the plurality of candidate ad words;
computing estimated upper and estimated lower marketing break even (MBE) values for each candidate ad word of the plurality of candidate ad words based on corresponding estimated upper and lower CPC values, wherein each estimated MBE value of a corresponding candidate ad word represents an estimated cost of an advertisement relating to the corresponding candidate ad word;
computing an average MBE value for each of the plurality of candidate ad words based on corresponding estimated upper and estimated lower MBE values;
selecting one or more ad words from the plurality of candidate ad words as a function of average MBE values associated with the selected one or more ad words;
receiving a measured upper and a measured lower CPC value for the one or more selected ad words based on an advertisement deployed relating to the one or more selected ad words, wherein the measured upper and the measured lower CPC values represent actual costs for one click on an advertisement related to the one or more selected ad words;
computing a measured average MBE value for the one or more selected ad words based on the measured upper and measured lower CPC values, wherein the measured MBE values for the one or more selected ad words represent actual cost of the advertisement relating to the one or more selected ad words; and
removing an ad word from the one or more selected ad words that has a measured average MBE value higher than a threshold.
8. The server of claim 7 , wherein the processor is configured for:
computing a global MBE average value based on the average MBE values for the plurality of candidate ad words, wherein the global MBE average value represents an average value computed across the plurality of candidate ad words; and
selecting a set of ad words from the plurality of candidate ad words that have their associated average MBE values below the global MBE value.
9. The server of claim 7 , wherein the selected one or more ad words include at least one of a positive ad word and a negative ad word, an advertisement based on the selected one or more ad words being displayed in response to the positive ad word, and prevented from being displayed in response to a negative ad word.
10. The server of claim 7 , wherein CPC estimates for each of the plurality of candidate ad words are provided by a traffic estimator based on number of clicks to an advertisement shown in response to a search query in a search engine having at least one of the plurality of candidate ad words.
11. The server of claim 7 , wherein the estimated upper and the estimated lower MBE values for each of the plurality of candidate ad words are computed based on one or more metrics from the group consisting of cost per click, clicks per unit time, conversion rate per click, conversion value, and cost per impression.
12. The server of claim 7 , wherein the plurality of candidate ad words includes keywords related to the seed topic and extracted from the information database.
13. A computer for selecting ad words for purchase comprising:
a graphical user interface; and
a processor configured for:
processing information from an information database using orthogonal corpus indexing and a seed topic to derive a plurality of candidate ad words;
receiving an estimated upper and an estimated lower cost per click (CPC) value for each of the plurality of candidate ad words;
computing estimated upper and estimated lower marketing break even (MBE) values for each candidate ad word of the plurality of candidate ad words based on corresponding estimated upper and lower CPC values, wherein each estimated MBE value of a corresponding candidate ad word represents an estimated cost of an advertisement relating to the corresponding candidate ad word;
computing an average MBE value for each of the plurality of candidate ad words based on corresponding estimated upper and estimated lower MBE values;
selecting one or more ad words from the plurality of candidate ad words as a function of average MBE values associated with the selected one or more ad words;
receiving a measured upper and a measured lower CPC value for the one or more selected ad words based on an advertisement deployed relating to the one or more selected ad words, wherein the measured upper and the measured lower CPC values represent actual costs for one click on an advertisement related to the one or more selected ad words;
computing a measured average MBE value for the one or more selected ad words based on the measured upper and measured lower CPC values, wherein the measured MBE values for the one or more selected ad words represent actual cost of the advertisement relating to the one or more selected ad words; and
removing an ad word from the one or more selected ad words that has a measured average MBE value higher than a threshold.
14. The computer of claim 13 , wherein the processor is configured for:
computing a global MBE average value based on the average MBE values for the plurality of candidate ad words, wherein the global MBE average value represents an average value computed across the plurality of candidate ad words; and
selecting a set of ad words from the plurality of candidate ad words that have their associated average MBE values below the global MBE value.
15. The computer of claim 13 , wherein the selected one or more ad words include at least one of a positive ad word and a negative ad word, an advertisement based on the selected one or more ad words being displayed in response to the positive ad word, and prevented from being displayed in response to a negative ad word.
16. The computer of claim 13 , wherein CPC estimates for each of the plurality of candidate ad words are provided by a traffic estimator based on number of clicks to an advertisement shown in response to a search query in a search engine having at least one of the plurality of candidate ad words.
17. The computer of claim 13 , wherein the estimated upper and the estimated lower MBE values for each of the plurality of candidate ad words are computed based on one or more metrics from the group consisting of cost per click, clicks per unit time, conversion rate per click, conversion value, and cost per impression.
18. The computer of claim 13 , wherein the plurality of candidate ad words includes keywords related to the seed topic and extracted from the information database.