I talk to brands around the world every day. Big brands. Multibillion-dollar brands. Often it is in the context of some adtech or martech business challenge. One of my first questions is usually: “Where’s the data?”
Today, most brands have contractual ownership of their data. What sort of data?
- Data used to target and execute campaigns, like first-party data
- Data collected during the execution of a campaign from ad tech platforms like an ad server, verification provider, DSP, research vendor, etc.
In the past, much of this data was owned by the ad agency of record. Other data sets might have been owned by the publisher or media owner. Today, it is a common contractual practice to have the entity cutting all the checks (the brand) remain the owner of all the data. Although, that ownership shift took the better part of a decade to achieve.
Contractual ownership ≠ control.
After I confirm that the brand owns the data, I typically say, “Ok great! Can we have a look?”
That’s where things get awkward.
There’s often a chuckle or some darting stares across the conference room table (back when in-person meetings were a thing) followed by, “Oh…uh…you’ll have to ask the agency.”
Just because you have a contract that says you technically own your marketing data does not mean you control it. It definitely does not mean you are getting value from it.
If you do not have explicit ownership of your brand’s campaign data in all of your contracts…OMG. What are you waiting for?!
Just last week, the ANA released new Master Services Agreement templates (free access) and resources for ad tech and martech procurement. Also, the ANA standard contract templates for brands and agencies include language to ensure proper data ownership.
Control ≠ value
Control of your brand’s data does not mean that you are getting a single penny of value from it. The data exhaust coming off of all your campaigns is one of the most valuable assets your marketing department has. Is all that data goodness sitting in a virtual drawer, collecting dust? Probably.
The primary reasons brands do not focus more time, energy, and resource on managing marketing data are:
- It isn’t sexy. Nobody gets invited to the VMA’s or the yachts in Cannes because they used the brand’s performance data to create a bespoke MTA model.
- It is nerdy. Very, very nerdy. First, there’s the architecture, then there’s the data science. Woof.
- It is a lot of work. It ain’t easy. There are 100+ moving parts and just as many variables. Plus, when you are just getting started, you will be wrong more often than you are right.
- Colleagues don’t get it. Getting your coworkers, managers, and executive excited by the opportunity is a beast. The entire time you’re talking, they are thinking, “What’s in it for me, nerd?”
Why should you care?
Remember when digital or new media was a cute little gig given to the intern to handle? No one took it seriously. Marketing crumbs were thrown at it just so the CMO could claim they were innovative. Now, the digital intern is the CMO.
That’s where we are right now with marketing data. The individuals that crack the code and generate millions of dollars in value using data that’s otherwise just sitting around will become the marketing gods of the decade.
Where to start
In other newsletters, I’ve discussed the critical importance of first-party customer data. Owning, controlling, and creating value from marketing data includes the first-party goodies and all of the other data exhaust that comes from your campaigns. Those are the tea leaves, so to speak.
For the purposes of this discussion, we will set first-party data assets aside and focus on data resulting from in-market communications.
Our goal for today is to gather as much of your marketing data as possible, get control of the source, and apply a base level of reporting capability on top of it. This is the first, most basic step to get you started, but expect it to take at least six months if you also have a day job with other responsibilities.
Time to hunt
To accomplish this, you need to go hunting. A full-blown, wide-scale audit is in order. You need to turn over every stone internally, at your agencies, and within your directly owned tech platforms to get a full picture of what is out there. This is a prerequisite for all data management related work in the future. The goal is to generate an inventory or manifest of all campaign data and related brand data.
Next, you need to understand what each data set contains, how it is constructed, and how the data is organized. To do this, you’ll need to roll up your sleeves and get your hands dirty.
For example, do you know what data your ad server collects? I promise you it is more than impressions, clicks, and conversions. Most ad servers have a row in a log or database for each impression with about 100 attributes and metrics for each. Your job is to understand what metrics are currently used versus what might be valuable to explore going forward.
Also, you need to get a sense of how portable the data is. Can you easily move it to an external reporting tool? Is there an API? Are there data logs?
Pools, lakes, and warehouses
Once you have a sense of what data is available, where it currently lives, and have thought up a couple of use cases, you will need to begin joining the data sets. Here is where new vendors come into the picture. You will likely need to create what is broadly known as a data lake. Note: This is where you’ll probably need to phone a friend. You’ll need a data architecture subject matter expert if you do not already have one.
Unless you are a quant or have some level of data science experience, this area will be very foreign. That’s ok. Just know that this is the central collection and processing point for all of your marketing data. It is where all the raw data finds its way home to you, where it belongs.
Show me the m̶o̶n̶e̶y̶ numbers
One of the most basic services built on top of your snazzy new data lake is a reporting service. The reporting service gets presented to you, the end-user, in the form of a data visualization like those generated by Business Intelligence tools (BI tools).
Some example vendors providing these services include (not exhaustive):
Note: I do not endorse any of these vendors. This is just a sample of the more popular choices you’ll find in the market.
Once you have your source data in hand, explorable and extractable by a basic BI tool, your mind will be unlocked. If done correctly, you’ll never go back to the dark days of offline reporting again.
Powerpoint reports are for amateurs
If the main way you consume or experience your campaign results is via Powerpoint reports created by someone else, stop it. Reports delivered to you via PPT, PDF, XLS, whiteboard, or crayon should be a thing of the past. The only appropriate use of Powerpoint for the delivery of campaign performance results is on a giant screen in a conference room or on stage at a trade show.
Most of your campaign analysis needs (85% +) can and should be delivered in a live BI tool of some sort. Why?
- Reports create questions. You can’t drill down in a PPT or PDF.
- BI tools and dashboards are live. The data in the reports are regularly updated, often daily. But certainly more often than a manual report.
- Checking a BI dashboard once or twice a week creates accountability.
- Small problems get solved before they become huge problems. If you have access to your data in near-ish real-time, your campaigns are less likely to go off the rails without you knowing.
In short, you should have direct, near real-time access to the totality of your campaign data, including ownership and control of the underlying source data.
Straight talk: this is hard work. However, it is 110% worth it and will be table stakes sooner than you realize. That means you’ll have the strategic edge over your competitors (and your colleagues).
Run, don’t walk towards completing these five steps:
- Do conduct an audit. Go hunting for all your data goodness. Understand where it lives, who controls it, and how to get your hands on it.
- Do explore creating a data lake. You’re going to need a place to store your data spoils. There are lots of vendors and technology solutions, but all require a nontrivial investment.
- Do select a BI tool. With or without a data lake (some BI tools have this capability built-in) you are going to need a way to experience, query, and visualize your data.
- Do get educated. Go ahead and spend a silly amount of time on YouTube. There are countless videos on this topic.
- Do remain the squeaky wheel. You are going to run into colleagues, managers, and others who just don’t get it. You’ll be told it is too hard, it takes too long, and it is too expensive. They’re wrong.
🥇Ari wrote the most read AdExchanger post of 2020. Adios cookies.
💸 Adweek had a busy week buying Social Media Week.
🚨 And maybe don’t illegally storm the US Capitol like the CEO of Cogensia did and expect to be in the mainstream marketing business afterward…