Arming the Rebels is a four-part series with Anthony Rotio, co-founder and co-CEO of GrowthLoop, for every retail media network that isn't Amazon or Walmart. The series is brought to you courtesy of GrowthLoop.


When I ran my agency, Bobsled Marketing, I was frequently tempted to build our own software: reporting dashboards, review scrapers, content optimization tools. Sometimes the tool we needed didn't exist off the shelf. Sometimes we could see a problem clearly that nobody in the market had solved yet, and owning the fix looked like an edge.

Each time our leadership team discussed these exciting possibilities, cooler heads thankfully prevailed.

Companies with entire engineering and product teams have the same problem: their eyes are bigger than their stomachs. Anthony Rotio, co-founder and co-CEO of GrowthLoop, told me about a conversation with someone who leads an AI team at a large organization. Their hand-wavy verdict on Anthony's company: "I could build GrowthLoop in two weeks."

He also told me about one of the top-tier retail media networks, which prides itself on building everything in its media business itself, including its own audience platform. The problem now is that nobody there wants to maintain it, and thus the RMN is now shopping for partners to take over some of the pieces.

In Part 1 of this series, we talked about what's holy: Once a network knows what's holy, it still has to decide what to build, what to buy and what to rent.

Gopuff knows how to answer that question. It handed its onsite ad technology to Instacart and kept the large custom brand activations for its own team. When I interviewed Michael Peroutka, Gopuff's head of ads and strategic partnerships, he said, "I'm a realist. I understand where we kind of sit on the food chain, and we do need to fight for every dollar." Gopuff also builds its advertiser audiences with GrowthLoop. So it rents one piece of the stack, buys another, and keeps what only Gopuff can do.

This interview has been edited for length and clarity. You can listen to the full interview on the podcast feed or watch on YouTube:

"Could we build this ourselves?"

Kiri: You must hear this all the time from retailers: "Could we build this ourselves?" How do you handle it?

Anthony: I actually love when this question comes up. We've been working with enterprises since day one. Google was our first customer, about a decade ago. When you work with customers that are that build-forward and innovative, of course they're gonna wanna build things. The principle has never changed. Build the things that will be differentiated for you, and a core competency that you want to own. And the GrowthLoop system won't be a black box where we're storing things. It's all going back into your own data layer.

Right now you can build a lot of things really quickly. So when you have this gluttony of riches, that's where strategy becomes really important. You pick: "We wanna invest in our data layer, because that consumer relationship is really important to us." Or, "We wanna own the last-mile delivery." Or, "We wanna own the big brand activations on site, because it's something only we can do as Gopuff."

We just released something called RMN in a Box, and there are a lot of things in the box. The whole point is you don't have to take the whole box.

Kiri: Scoping a problem and building the fix must be satisfying for an engineer. Maintaining it and integrating it is a whole other workload.

Anthony: There's a saying on the manufacturing and production side: demos are easy, scale production is hard. Maybe you could build a demo that looks a little bit like GrowthLoop in two weeks. But to actually serve thousands of audiences at 2,000 requests per second on an onsite or personalization experience, these things are much harder to do long term. Some organizations want that to be their muscle, and they hire up for it.

But invest in your customers, invest in your products and services, invest in the things that differentiate you as a business, rather than technology that you can rent or work with a partner on. And if one day you do decide to build it, don't you wanna learn along the way from folks who have done this?

Clean rooms

Kiri: Data clean rooms are getting a lot of attention, and some retailers are uneasy about sending personal data to a third party. GrowthLoop has a composable clean room that lets retailers avoid doing that. Why did you put it in the stack?

Anthony: The first question we ask is, is this a technology challenge or a logistics challenge? Nine times out of 10 it's logistics: "I have to go through vendor onboarding and all my internal processes to let other people have access to this clean room." So we've tried to make it much easier to do the overlap between the audiences the brand already has and the retailer's, or to exclude people you're already targeting, in a way that preserves the policies of both organizations. It's never sending raw PII downstream.

The maintenance bill

The decision to build gets made in a meeting held to discuss the internal business case. But the maintenance lingers on for years, and at that top-tier network, nobody wanted to own it.

Renting carries its own risk. In Part 2, Anthony warned against building up customer learnings "in a system that relies on a contract that may not be in existence in two, three, four years down the road." Whether a network builds or rents, the test he applies to his own product is the useful one: does the data, and what the network learns from it, end up in the retailer's own warehouse? Vendors find that easy to promise and buyers find it hard to check, so ask any partner, GrowthLoop included, to show you where the learnings land.

The case for renting gets stronger as advertisers ask for more. In GrowthLoop's Commerce Media Investment Index, 95% of advertisers say access to more audience segments would make them more likely to increase spend with a network, but only 24% have access to 100 or more segments across their partners. A homegrown stack sized for today's requests has to keep growing to meet that. And 93% say they'd be more likely to invest in a network that could activate audiences without exposing personal data.

Every network will draw the line between build, buy and rent somewhere different. The important part is deciding what's truly essential to keep for yourself.


💡
Standing up a network, or rethinking the one you have? GrowthLoop's RMN in a Box is a set of tools for running a retail media network, and you don't have to take the whole box. Learn more

More in the 'Arming the Rebels' series:

  1. What’s Holy? How much of the customer relationship a retailer should let brands use.
  2. Martech Before Media. Why the first job for many new retail media leaders is fixing the marketing team’s customer data. [LINK]
  3. Build, Buy or Rent. You’re reading it 😄
  4. Speed Beats Size (coming Thursday). Why a smaller network’s best shot at the biggest budgets is getting audiences live, and results back, fast.