Ad servers have direct access to every ad call that is made to them and therefore have a very precise measure of the ad deliveries that they are responsible for. This means that, as a best practice, you should always use ad server reporting for delivery volume metrics, like number of served impressions, or clicks, conversions, and by extension, rate numbers, like CTR and conversion rate.
Lotame, on the other hand, is measuring an entirely different kind of metric with Lotame Insights: the profile of users on a campaign rather than the number of users. Lotame captures browser cookie IDs and mobile IDs that have interacted with a campaign and compare those to Lotame’s entire profile store. The Lotame Insights metric “Tracked Unique Impressions” is not how many unique impressions were served by your ad server, but rather as a measure of context for how many measurable profiles were tracked on this campaign to perform overlap calculations.
Differences between the two kinds of measurements
Ad servers can count each time a request was made from a web page to display an ad stored on the server, how many times media was delivered to and loaded by the page, and how many times a click activity fired a request for new media or a redirect. All of these methods are a direct request to the ad server.
In order for Lotame to capture the profile ID of a user who interacts with an ad, we do one of two things: 1) Lotame client places an image pixel on their creative media, which, when loaded onto a page, grabs the profile ID from the browser, and sends it to Lotame’s servers to be matched with the campaign. 2) Or, Lotame client passes into Lotame the ad server logs. In both cases, Lotame is not a direct party to the ad call, therefore relying on the third-party partner to fire or execute Lotame’s code in order to capture our profile ID. Lotame takes great care to only record an event if we have been passed a profile ID correctly. Lotame is interested in getting the precise profile ID for users.
What Lotame does with the data we capture
Once Lotame has captured a valid profile ID from a campaign interaction, a data point is appended to that profile ID that says “This profile ID did this campaign interaction.” The unifying hub of our data storage technique is the profile ID. Lotame takes all the profile IDs that have that same data point and performs overlap calculations for that group. Overlap calculations are the process of counting how many profiles share the same data points and then grouping them and seeing how each group overlaps with the other.
Let’s take “Cat Lovers” and “Dog Lovers” as an example. Lotame counts how many of our billions of stored profile IDs have the data point “Cat Lover” appended to the profile (i.e., 200MM). Then we count how many profile IDs have the data point “Dog Lover” appended to their profile (i.e., also 200MM). Next, Lotame counts how many profile IDs have both “Cat Lover” and “Dog Lover” appended to the profile. 50MM is the outcome. Lotame calculates that 50MM is 25% of 200MM and reports that “Cat Lovers” have a 25% overlap with “Dog Lovers.”
Lotame clients use this overlap data to find groups of users that closely match the kinds of users they are trying to reach.
It’s important to point out that in order for Lotame to report accurate overlap calculations for “Cat Lovers” Lotame does not need to capture every user who has ever loved a cat. Instead, a mathematically significant number is needed in order to calculate overlap with other data points.
Now that you know the differences between the way Lotame and ad servers capture and calculate campaign data, let’s review two ways that you can leverage these differences.
Discover New Avails Precisely
You can use Lotame Insights to discover segments that are highly likely to perform for a particular campaign and then use your ad server reporting to get a precise sense of the available inventory for that new segment.
Contextualize and Profile Engaged Users
Take the precise numbers of an interaction from your ad server, (i.e., 3,000 clicks) and use Lotame Insights to build a behavioral profile of that group of users (i.e., Compact Car Intenders, Urban, Employed).