Previously, we shared part I of our interview with our Director of Product, Stuart Moncada, which covered the industry at large. This is part II, where we focus on the programmatic space, including partner relationships, challenges, best practices, and insights.
If there is one thing every ad ops team should do at this moment in the industry, what is it?
Hold your partners accountable. Understand the value provided by each partner, understand the data, and set the bar high for transparency in the data they provide. This may even mean that you need to demand more visibility from your partners. They have the data, there's just a reluctance and a lack of urgency to make that data available and to provide the right tools.
What should organizations focus on when evaluating potential vendors?
I think there's several things to focus on, but monetization comes first. Publishers will want to look at things such as revenue and fill rate—how much money they're going to get and how much will they be selling. What is each partner’s contribution? If someone is selling half of your inventory, that's a very important partner to you. If, on top of that, they’re doing it at a very high CPM, then they have a big part of your business.
But I think publishers need to also carefully weigh other aspects of their demand partners, like transparency:
- How open are they about their fees?
- Do they tell you specifically what their rev share is?
- Variable or fixed?
- Is it fair?
- Are they being open with the data?
- Are they really giving you all the data that's available?
Because there are certain partners right now that don't tell you which advertiser bought the inventory.
And then, what are the partner’s operational capabilities? Some partners have robust APIs and provide real-time data. Then there's other partners that you’ll have to manually pull reports from and you can’t set up email or report scheduling.
Again, the primary consideration will be the revenue side. But I think the operational side tends to have a much larger effect than people generally anticipate—especially when you get up to 10 or more partners. If you only have one or two partners, it's not unmanageable, you can manually deal with operational inefficiencies. But as you're trying to scale, you need more processes automated; and this is where you really become limited by the capabilities of your partners.
What is the biggest challenge to optimizing programmatic operations?
Well, the two big factors are fragmentation and a lack of standardization in the industry, which are correlated. The fragmentation has led to a lack of standardization because everyone has to find their own conventions, their own standards, their own scales for everything.
So the biggest challenge becomes how to cut through all of the noise (from what can seem like an overwhelming amount of data) to figure out what to measure, and how to measure it, in order to make intelligent decisions. And if you're not doing that, you're just flying blind and probably leaving a lot of money on the table.
What are some best practices for publishers in the programmatic space?
There are a few publishers out there who have made a very large investment in understanding and organizing their data by implementing a platform like Ad-Juster's or a homegrown solution. This includes automated systems to ingest the data. They aggregate and normalize it so they can compare apples to apples as much as possible. Next, they define baselines, standards, and KPIs that are important to their business; and then use dashboards and reports to track to those KPIs and standards.
Once that’s in place, publishers can start figuring out how to layer data science on top. For instance, instead of just having simple thresholds on revenue and filters, publishers could have dynamic algorithms that update their price floors automatically. And, they could prioritize the inventory they’re sending to certain demand partners to really optimize the revenue they’ll get for their inventory.
“Normalization” comes up quite a bit, can you break that term down a little more?
Normalization, sure. I feel like it's a bit ambiguous and people use it differently and for different things. In the context of programmatic data reporting, the term usually refers to 2 things:
Column mapping, or normalizing your different data sources so that you're mapping the right column from partner A to the right column from partner B. So if Google calls this column “X,” and Rubicon calls it “Y,” normalizing it so that you have those those columns stacked on top of each other or within the same global column.
I've also often heard normalization used to refer to matching values within a single column, such as the advertiser column, where this is very problematic. Continuing with the Advertiser column example, Google will call the automaker Ford, "Ford" while Rubicon will call Ford, "Ford Motor Company." These two data points need to be normalized so that they are treated as the same data point. And this is a frustrating problem for publishers to solve because it’s very time-intensive to keep up with.
And lastly, what are some unique insights publishers can glean from their programmatic data?
There's so much data involved with programmatic and you have so many players in the value chain that you really can uncover a lot of insights in the data. For example, you can get a sense for who the DSP, advertisers, and/or agencies are that are buying certain types of inventory. You can calculate average CPMs across different channels and formats.
It depends on what you’re interested in and where you see the biggest revenue-driving opportunities. You can get information on on bid metrics and figure out who's bidding on certain impressions. You can find out if you have redundant buyers across your different demand sources—for instance, a publisher should be able to help connect the dots for an advertiser, recognize buying patterns, and the create the optimal ad product so the advertiser hits their goals faster and the publisher improves their yield.
More and more ad dollars are transacted programmatically and publishers’ ad ops must adapt to support both direct-sold and all variations of programmatic. The data landscape continues to grow in size and complexity, but that shouldn’t be a cause for concern. Within all that data is a goldmine of information and insights. The right foundation and a trusted partner, like Ad-Juster, will help you turn that data into knowledge, awareness, and profit.
That wraps up our interview with Stuart Moncada, Ad-Juster’s Director of Product. What topics piqued your interest? Do you have any questions for Stuart? Contact us and we’ll include your questions in the next interview.