Podtrac admits to error on website as it aims for IAB Certification
This article is at least a year old
Last month, podcast analytics company Podtrac’s overall download figures appeared to decrease, as the company made an update to its measurement algorithm as part of a change related to IAB Certification work.
We noted a drop for downloads, overall, of about 20% in the Podtrac published charts; though we hear that some publishers are seeing drops in their Podtrac download figures of almost 50%: placing them in a difficult position if they used these numbers for ad sales.
In an FAQ sent privately to publishers but seen by Podnews, Podtrac said that they’ve made two major changes: “We changed our lookback window to be more conservative” and “removed partial downloads”.
Last June, Podtrac were bullishly claiming that they were “IAB compliant… and then some”. “Podtrac audience counts are, and have always been, consistent with the best practices outlined in the IAB Podcast Metrics 2.0 document.” The document said “Podtrac uses a 24-hour look-up window”, and that they “use advanced algorithms for partially downloaded content”.
A “window” is recommended in the IAB Podcast Measurement Guidelines v2 as a way of avoiding double-counting downloads. If the same file is being downloaded multiple times by something that looks like the same device within the same time window, it should be treated as one download, say the guidelines.
However, Podtrac’s page was quietly updated earlier this year. Mention of both the 24-hour window, and of partially downloaded content, were both removed.
We asked Robert Freeland, CTO of Podtrac, to clarify the changes that were made to the 24-hour window. He admitted, in an email to us:
Our original post should not have mentioned a 24-hour lookback window, because that was still under development and not yet live.
He added:
The specific IAB windowing guidelines are very clear about not mandating a 24-hour static window, stating: “shorter windows…should be done with care” and “companies are allowed to use more sophisticated mechanisms.” So using a 24-hour window is not necessary for compliance, and our older algorithm was still IAB compliant.
He also said that the page was updated in February “to focus less on the technical comparison of our latest vs. previous algorithms until our IAB certification is complete.”
We asked about what change the company has made with regard to partial downloads. Removing downloads with less than sixty seconds of audio has always been part of the IAB specification, so we were curious as to what had been changed. Robert Freeland responded:
We have used various algorithms to account for partially downloaded files over the years. We believe that those algorithms met the IAB’s specification: “If the logs do not have byte range request data, more advanced algorithms that factor in a correction for partially downloaded content may be used.” In our most recent update -- after significant research -- we developed a new mechanism for doing so that yields more conservative results.
On the changes to the page itself, he told us:
The “And Then Some” was marketing speak for the sophistication of our algorithms and the years we’ve been doing this, but we decided in February that it was perhaps too confrontational given all the sensitivities around the IAB certification process. Thus the toned-down version of the page that we’ve had live since February.
Podtrac’s IAB compliance work is ongoing, and as Robert Freeland says, “I think we’re the only company communicating our status during the process, which can sometimes be challenging.”
Early in 2018, Podnews discovered a serious error in Podtrac’s figures, which has since been corrected. Last year, we noted that Podtrac figures were at least 27% higher than statistics from Blubrry, which at the time was nearing IAB certification. But in recent months, the company has been realigning its work to be compliant with others.
A level playing field, for all, is important to help the industry grow. We welcome Podtrac’s efforts to achieve IAB Certification, and for Podtrac to return numbers similar to other podcast analytics companies. We look forward to reporting Podtrac’s figures with confidence in the future, and thank them for their willingness to engage with us.
