When to bring measurement in-house, and when SaaS is enough
Most advertisers are well served by measurement software. A specific few are not. Here are the signals that you should own it.
Most advertisers are well served by measurement software, and should be. A specific set are not, and for them, owning measurement is the right call rather than renting it. The hard part is knowing which group you are in.
Who should bring measurement in-house
In practice, the companies that benefit from building their own measurement share some combination of four traits.
- An intricate martech stack that does not connect cleanly to scaled SaaS tools.
- Business or marketing nuances unusual enough that they need bespoke analytics, not a template.
- A desire for full ownership of the data science: the model code, the testing program, the assumptions, and the roadmap.
- A shift already underway from legacy attribution toward genuinely scientific, causal measurement.
Think of the Netflix or Booking.com profile: businesses whose measurement problems are, in effect, one of one. Scaled tools were not built for their stack or their questions, and no amount of configuration closes that gap.
When SaaS is the right answer
If you do not fit that profile, buy. There is no prize for building what you could purchase. One caveat: only a handful of measurement platforms are genuinely excellent. Many look sophisticated until you lift the hood. If you are going to rent, evaluate the methodology, not the interface.
What we build
Growth By Science works with the companies that scaled tools cannot serve: fully bespoke, in-house measurement systems, and the one-of-one analytics problems that come with them. We have helped leading brands optimize between 10 million and more than a billion dollars in annual ad spend, using everything from incrementality testing and custom marketing mix models to causal machine learning, clean rooms, and measurement strategy.
Own your measurement.
Common questions
Should you build your own marketing measurement or buy a SaaS tool?
Most advertisers are well served by SaaS and should buy. Build when you have an intricate martech stack that does not connect to scaled tools, unique business nuances that need bespoke analytics, a need to own the data science itself, or a shift already underway from attribution to scientific measurement.
What kinds of companies bring measurement in-house?
Businesses whose measurement is effectively one of one, such as large streaming, travel, marketplace, or gaming brands. The Netflix or Booking.com profile is typical: the stack and the questions are unusual enough that scaled software does not fit.
Are all measurement SaaS platforms the same?
No. A handful are genuinely excellent; many look sophisticated until you examine the methodology. If you buy, evaluate the underlying approach rather than the interface.