Data Questions that Are Surprisingly Difficult for Publishers to Answer

Every publisher wants to be "data-driven". But every ad ops professional knows that being truly data-driven is harder in practice than in theory.?

To make a data-driven decision, ops pros often need to collect data manually across dozens of SSPs, ad servers, CRMs, OMSs, and other systems. Then they need to normalize the data. Only then can they analyze it and make a decision.

The upshot? There are many questions that are surprisingly difficult for ad ops professionals to answer. And the absence of answers makes it harder for ad ops pros to do their jobs.

Here are a few common questions that should be easier to answer — and how ad ops pros can get answers faster.

How much did a specific advertiser spend with us via the open programmatic market last quarter?

To answer how much a given advertiser spent via the open programmatic market in a quarter takes more than just completing a few summations. It involves navigating all the platforms that the advertiser may have gone through to buy inventory. This requires an ad ops professional to go through potentially dozens of SSPs — including, in some cases, sifting through a given SSP’s multiple separate instances.?

What’s more, you also have to account for and unify all the data records that refer to the same entity but are named differently. Let’s imagine an advertiser named “Tortoise and Egg.” Across platforms, the advertiser could appear as “T&E” in one place, “T_and_Egg” in another. Not to mention the names of the agencies or partners Tortoise and Egg buys inventory through.?

Publishers shouldn’t have to waste their time sorting through systems to get this information. Automation should aggregate and normalize it for them so that it’s ready and up to date whenever they need it.

How much money have we made this month on desktop?

Again, answering this question could require you to pull data from 20 different systems and then normalize all the data across devices and browsers, all while ensuring that extraneous data points like mobile revenue are separated out.?

In other words, chasing after desktop revenue?— much like answering how much an advertiser spent via open programmatic —? means endless Excel sheets and Vlookup on Vlookup on Vlookup.?

What is our revenue per pageview (RPM) on different sections of our site??

Another seemingly reasonable question, but to put the pieces of this puzzle together, you again have to assemble all the revenue data from each of the publisher’s revenue systems. Then you need to collect pageview data from a web analytics tools such as Google Analytics. On top of that, each system may keep track of pageviews differently, which can result in discrepancies that are tedious to normalize.??

Surfacing these kinds of data points —?open programmatic revenue, desktop revenue, RPM — takes hours of an ad ops pros’ day, both due to data overload and data complexity.?

Get Out of Vlookup Jail, and Make Faster Data-Driven Decisions

More revenue lies on the other side of readily available data. For example, publishers keep much less of each dollar advertisers spend with them via the programmatic open market relative to direct-sold inventory. If you know how much each advertiser is spending with you on the open market, that’s intelligence you can use to strategically approach the best candidates to upgrade to a private marketplace or even deeper relationship where each party keeps more of every transaction.?

But making these strategic decisions is impossible — or it takes much longer to do — if you’re doing lookups on lookups on lookups to collect each interesting data point. By using Burt, these data points and others no longer require manual data aggregation and normalization. All the data you need is at your fingertips — up to date, normalized, and accurate enough for your billing team to work off of. With Burt, you can take back control of your day and make the decisions that will make your CRO feel more confident when they say, “We’re a data-driven organization.”

If saving hours per day by getting better data faster sounds interesting, contact us.