It’s been eight months since Facebook “Inflate-Gate” (as a refresher, see admission from Facebook), where they miscalculated the average view time on video ads during a two-year period, but the saga continues. Earlier this week, Facebook admitted to their fifth metrics error and this time, they are refunding some advertisers.
Advertisers are getting frustrated and are starting to take things into their own hands. Adweek reported last month that Fiat Chrysler Automobiles, who spends over $1 billion in advertising, was fed up that Facebook’s data is not compatible with other media outlets, so they created their own set of measurement standards for video views and other additional stats.
Advertisers taking ownership of key functions and data is not unusual. Many have brought media buying in-house and some have implemented DMPs internally to give them control over their audience data. At Origami Logic, we foresee a trend in advertisers managing their marketing performance data in-house, but it won’t be easy.
Getting the metrics right is not easy. The amount of data being processed on a daily basis by platform providers like Google, Facebook, Twitter and such is overwhelming, and we have to remember that the people who are writing the code to calculate the metrics are human and all humans make mistakes.
Also, it is difficult to stay on top of all of the metrics “changes” being made by the platform providers fairly regularly. We categorize the “changes” in the following manner:
- Adjustments – There are certain metrics that are adjusted over time, for very valid, innocent reasons. This happens most often with regards to ad click rate metrics. The primary reason they are adjusted is because, over time, a platform provider may come to the conclusion that certain clicks were a result of spam (some spam cannot be identified instantly). Since the number of clicks an ad receives has an effect on spending, these type of adjustments typically happen – and should happen – right before the end of a billing cycle.We have also seen adjustments made to the daily total metric for website visits. While the total visits for a month do not typically change, we have seen the count for visits shift from one day to the next. We have yet to determine why this happens.
- Restatements – Platform providers do make mistakes in calculating certain metrics and have to restate the results of metrics like Facebook had to do with their video metric. As mentioned earlier, getting the metrics right is very difficult, particularly since some of the problems can be very subtle.
- Redefinitions – Some metrics are redefined by a platform provider, for various reasons. In some cases, after a metric is redefined, the results may be recalculated historically with the new definition and in some cases, the platform provider may just use the new definition and its results moving forward.
At Origami Logic, since we provide a marketing performance measurement platform that connects to a wide variety of data sources, like DoubleClick, Facebook, Google AdWords, Google Analytics, Instagram and Twitter, it is our business to understand marketing performance data. We spend a tremendous amount of time understanding each of the data sources we support and its data at a very detailed level. We also devote a lot of energy in trying to find inconsistencies and anomalies in the metrics we digest from the platform providers. In particular, there are a couple of things we do:
- We constantly check the latest results with historical data to find anomalies. We do this by not just grabbing the latest data and comparing it to the data we previously extracted. We are constantly back fetching data from a broad time window since certain metrics could have been adjusted over time, as described above. Most anomalies we find are a result of actions taken by a brand, like stopping a campaign, but some are not and could be caused by a platform issue.
- We calculate data from different perspectives to ensure consistency. We fetch data along a variety of dimensions (account, campaign, keyword, etc.) and we make sure that all related calculations are consistent. Sometimes, we have to use different APIs from one platform provider to access all of the metrics needed and it is not unusual to discover inconsistencies between the APIs.
As we mentioned earlier, getting the metrics right is not easy, even for the large platform providers. Therefore, it is important for marketers to have a strong “data partner,” somebody who knows the data well.
Understanding marketing data is at the core of our business. With an ever-growing list of marketing data source integrations, the Origami Logic platform gives brands the ability to converge marketing performance data across different technology stacks and receive the transparency they need to truly understand how their campaigns are performing. And furthermore, we just announced our newest capability, Workbench, where analysts get direct access to raw marketing performance data. How amazing is that?
We welcome the opportunity to partner with brands and agencies and help them navigate their way through this complex – and important – journey. If you would like to get started with the Origami Logic platform or participate in our Workbench beta — get in touch!