"Brand Aware” explores the data-driven digital ad ecosystem from the marketer's point of view.
Today's column is written by Jayne Pimentel, senior director of growth marketing at DraftKings.
One of the more far-out ideas to emerge in the post-digital era is the simulation argument, which stipulates that we are likely living in an advanced computer-generated reality. The hypothesis is a fun joke to some and a serious worldview to others.
But for data- and technology-forward marketers, the simulation can be more than a thought experiment. It is a reality, but one in which marketers are the ones running the simulation, while vendors, partners and platforms are living in it.
What follows are some thoughts on how to be a good steward of your simulated reality and set up rules that make it more utopia than dystopia.
The simulation argument takes on greater applicability in today’s digital advertising industry with the rise of in-housing.
In-housing empowers advertisers to have a one-to-one relationship with a brand’s data just like they do with their customers. This enables every brand to make its own simulation where each vendor resides.
As a brand, if I want to add a multiplier to my event signal, for example, I’ll do it. If I want to adjust the difficulty of my metrics for a publisher or platform, like going from normal to hardcore mode in “Grand Theft Auto,” then the simulation also changes. The ownership and control granted via in-housing creates a brand ecosystem with more easily refined data and accountability.
In my simulation, predictive models reflect the data I want them to use and enable me to show partners what I want them to see.
Predictive models are great! The majority of our digital marketing efforts at DraftKings are focused on some form of predictive modeling.
Typically, it’s based on a platform’s models (people-based and action-based) or our own proprietary models for channels like social and programmatic. I imagine we’re part of the norm, where most of our budgets are allocated based on the impact of predictive models.
Due to our reliance on predictive models, our media investments have been de-risked, but they have also become less tangible for a partner. In other words, what a partner sees is what I want them to see.
Similarly, a predictive model is only as good as the data it receives, so it too sees what I want it to see. Now you may be thinking I’m about to consider myself an omnipotent demi-god in this scenario, but that is far from the truth. I'm more of a narrator in this simulation, which is like the advertising version of “Westworld.”
In my simulation, advertisers are or will become their own attribution providers.
We read headlines lately about the end of cookies, platforms building higher walls in their walled gardens and a world without device IDs.
I don’t lose sleep at night. Most sophisticated advertisers have planned for that day and stockpiled their databases accordingly. They’ve also made efforts to create registration and sign-up events to ensure persistent identifiers to track.
When brands become their own attribution providers, it is inevitable that their perception will become the reality for all platforms and publishers. In fact, the more platforms pitch me on why we should move all our attribution to them, the more I pray for the speedy end of days where advertisers decide to own their attribution.
Today, I have to manage conversations with players in three different realities: self-attributed networks or platform-based attribution; mobile measurement partner/third-party attribution; and DraftKings/advertiser-based final attribution. As platforms share less data they become less relevant in a world where advertisers are making the attribution rules and shifting to incrementality rather than last click.
Onus of the marketers
Now I’d like to put the pressure back on marketers. With the decisions we make to change our simulated environment come the responsibility to treat this simulation like your platform/publisher/partner’s reality.
Moreover, no one will want to play, let alone survive, in your simulation if the rules are stacked, the sweat equity is ill-balanced, and your story mode is one dimensional. For instance, if you give a partner like Facebook all your data – revenue, mom’s maiden name, favorite color – then you better be willing to do the same for every other platform and partner in your network.
It is not fair to arm one player with unlimited lives and a bazooka while another is blindfolded and barehanded – marketers need to do better to create balance and parity in their simulations.
Additionally, think about the basic human needs of your partners:
- How does a partner live in my simulation?
- How does a partner achieve longevity?
- How does a partner achieve being liked/loved?
- How does a partner create influence over the simulation they’re existing in?
Steps to simulation success
The answers to the above are all fundamental actions. Be upfront with a partner and provide concrete expectations. Fluff metrics benefit no one, and I find proxies useless. Do not waste a partner’s time by telling them they have to hit a 75% completed view rate when your actual metric is revenue. Do not burn through partners, but instead tighten the screws on a few core ones.
Make them sweat. Make them work. But work with them. This goes back to predictive models. Most models take time to fine-tune, and that goes for your proprietary algorithms as well as a partner’s, so you need to give them time and appropriate data.
A basic human need is to be well-liked and loved. A partner needs the same. To invoke this as a marketer is to get partners to perform. If you do a good job at giving partners the data they need to succeed, set the right difficulty setting and appropriate sweat equity, then give them praise for a job well done.
Lastly, if a partner wants influence in my simulation:
- Never lie to me. I know your player attributes, so don’t come off as a 99 in shooting accuracy when you’re a 68.
- Go through story mode in my simulation, which means don’t go after my largest business problems or pitch me that your lifetime-value measurement product is better than my internal attribution without being able to recite our internal methodology from memory.
- Find the Easter eggs in my simulation – show me something I overlooked or did not consider before.
As you can see, a fair and reciprocal relationship is fundamental for success in any simulation. Hopefully you found this informative, maybe even got a few laughs too, and if it all seems strangely familiar, don’t worry – déjà vu is just a glitch in the matrix.