Frequently Asked Questions

Terminology

What is a query?

A query is a search that someone typed into Google, Yahoo!, Bing, or one of the other search engines to get to a site.

What are patterns?

Patterns are searches that have a similar structure. For instance, the pattern “jobs in X” represents searches for jobs in some location. Concentrate automatically discovers patterns in search queries.


Loading lists

What types of files can I upload?

You can load Excel, comma delimited (CSV) or tab delimited files. Concentrate recognizes query lists from a variety of leading search analytics and intelligence solutions including Hitwise, Comscore, Trellian, Compete.com and Google Analytics.

How long will it take to process my list?

It takes about one minute per one thousand queries to process a list. So, if you have a list of 30,000 queries, expect Concentrate to take about 1/2 hour to find the patterns in that list. For long lists, you can have Concentrate send you an email when the list is complete.


Query patterns

Does Concentrate just look for predefined patterns, or discover them from the data itself?

Concentrate discovers patterns in the queries themselves by looking for queries that share similar word patterns.

Are the patterns completely automatically generated, or is there a human component that goes through and selects only the best templates?

The patterns are completely automatically generated.

How are queries assigned to patterns?

Queries are currently assigned to the most specific pattern that they match. Some patterns are very general (like “jobs [x]”), while others are quite specific (like “things to do in [x]”).

What does the "X" mean in patterns?

The "X" is a wildcard that can stand for one or more words. These "masked terms" are often variants of similar concepts, like locations or celebrity names. Over time, Concentrate’s learning algorithms will replace many of these wildcards with named entity labels. For example: "hotels in X" might become "hotels in [City]".

Why are some queries "Uncategorized"?

There will always be some queries that Concentrate can’t fit into a pattern. The percentage of total queries that are Uncategorized is affected by the following factors:

  • Many of the uncategorized queries are single term queries or rare phrases that occur infrequently. Many very broad queries will not fit a phrase pattern ( for example, single terms like "flight" or "hotel").
  • Concentrate finds better patterns if it has more queries to work with. If you load a list containing just a few hundred queries, fewer patterns will be found.
  • Concentrate finds more patterns with targeted sites. A site focused on a single industry or kind of user need, like expedia.com will have fewer uncategorized queries than a broad site like nytimes.com.
  • Many brand related terms are not currently recognized by Concentrate. As we add more named entity lists, these patterns will be found and uncategorized queries will decline.


Analyzing results

How can I find queries that only contain a specific word or phrase?

When you are using the Query Explorer, you can filter queries using the filter box below the Query Explorer. This will limit queries to those containing the exact words you type. If you type more than one word, only queries containing that exact phrase will be found. For instance, if you filter on "cheap hotel", you will match "cheap hotel in new york" but you will not match "hotels that are cheap".

How can I download the results from Concentrate?

You can find a "Download" link just below the sites name. This will download a comma-delimited list of all the queries in the list along with the pattern each query belongs to.


Limits and access

How many lists can I process per month?

Concentrate does not limit the number of lists you can process. However, we do reserve the right to throttle performance in order to ensure a good experience for all our users.

What is the maximum length of a query and maximum number of queries that can be processed in Concentrate?

Query length is currently limited to 300 characters (about 40 words). The maximum number of queries that can be processed depends on the level of your plan, which can be found here. If you upload a list containing more than the number of queries paid for in the plan, only part of the queries will be used.

Does Concentrate offer an API?

Concentrate plans to allow direct API access to the pattern-discovery algorithm which will allow the integration of the Concentrate discovered patterns in your business processes. For more information, contact Concentrate support.