Required Skills: Clairvoyance, (Human) Computing, & Culture

by Laura Boodram

Recently, FatTail and Beeler.Tech gathered publishers of all sizes together for a conversation about the platforms they were using to make inventory forecasting and yield management easier. After an hour conversation, we found that there were no software unicorns out there that solved all of their issues.  What we did learn, though, was that the best forecasting solutions have far less to do with technology and far more to do with the people behind the process.

Whether you have a dedicated role for predictive analysis or whether it’s bundled with revenue / ad operations / yield and inventory management / audience development / janitor, someone inevitably ends up in the role of the clairvoyant - that is, the person who will predict what will happen against hurdles like:

  • Traffic that fluctuates based on news events or sports outcomes. “I basically have a model for winning the world series, going to playoffs, or getting knocked out early in the season.”

  • Segmented inventory sold based on traffic or region with traffic variations

  • Last-minute deals that change expected delivery on everything that’s already in flight. “We have to go line by line on existing campaigns, check forecasting and results, adjust as needed. It’s really hard as it requires you to understand the site and network before you understand your job as an operations person.”

  • New product or site launches. “Product wants revenue from day one on something new that’s launching that you have no benchmarks against.”

  • Content marketing/SEM that causes unnatural spikes in traffic that are hard to adjust. “Site developers helped implement where a key value is passed based on a url string and so we can segment and report on organic inventory. Still get the full picture but can slice organic versus non organic and build that into seasonality modeling.” 

We also heard several examples of siloed decisions to make changes to the site or ad structure without consideration for the revenue impact - only to discover the issue after weeks of lost revenue.

To remediate these types of issues, some publishers are ingesting site level traffic data and looking at dashboards that show day-over-day / week-over-week fluctuations that allow them to hone in on abnormal trends and identify if there is reason for concern.

The overall goal is more automation around EFFECTS and then it’s really the human’s job to figure out the CAUSE. 

As we wrapped up the call, we asked one final question: 

“What’s the one thing you would change about people, process or platforms if you could?” 

And even with all the best tech and processes in the world, the consensus was to solve for the human element.

If you end up in a situation where inventory was not what you expected, you go through the motions to identify why, and whether a particular person / team / technical issue is to blame. Nine times out of 10, a closer loop between product releases that impact ad inventory, back to the development organization turns out to be the most helpful thing. 

The group agreed that a tighter feedback loop - where everyone understands the shared goal of maximizing revenue - is usually the best way to help reduce issues moving forward.

More / different to share? Want to continue the discussion? Head over to the Beeler.Tech publisher workspace. Not a member? What are you waiting for?! Contact Rob directly to learn more.

Previous
Previous

Salesforce<>AdBook+ Integration Unites Sales & Operations Galaxies

Next
Next

4 Tips for Getting the Most Bang for Your Revenue Management System Buck