Retire the 500-Slide Plan: Test-and-Learn Operating Models and Smart Partnerships

This is the last of five posts from the Fierce Pharma panel on the next era of marketing growth. It’s about the operating model underneath everything in the first four, from personalization to inverting the 70/30.

Static planning is the first thing to go

One panelist described the familiar annual ritual: a 500-slide deck, built 18 months before the year it governs. Strategy shouldn’t change every day, they said, but static planning needs to give way to funding learning and reacting to what you see, because change is arriving faster than ever.

What has to be true for that to work? The panel’s answer was cultural:

• Reward curiosity and let people make decisions without penalty.
• Create safety and empowerment while keeping goals and outcomes clear.
• Normalize test and learn. One speaker closed the session by saying teams are good at launching and weaker at testing, and that fear of failure discourages the experimentation that produces learning. (Jeremy MacDonald made the same point at BrightonSEO with a rocket: finding failure is success.)

Don’t build on a fixed stack

A second theme was the risk of building on a fixed technology stack. A panelist described the pattern: you spend six months building, and by the time it ships something better has arrived. The recommendation was to keep people, process and tools flexible, and to avoid over-engineering rigid pipelines. Their summary: focus on those three, and build tools that can be swapped in and out.

That advice lines up with how fast the tooling is changing across marketing. If a workflow is documented, modular and tested, replacing one component is a small change. If it’s a monolith, it’s a rebuild. Tools like n8n make that kind of modular workflow easier to build.

Use partnerships to close capability gaps

The panel’s final operating-model point was partnerships. In rare disease, where finding patients and accessing data are hard, one panelist said they look for unique partnerships that provide competitive advantage. Their example was a San Francisco company whose patients volunteer their electronic medical records and genetic data, giving researchers and pharma a database of ultra-rare populations, with an AI agent layered over the data to coach patients. The point was that sometimes the solution doesn’t have to come from inside your own organization, and that the right partner builds trust.

What to prioritize in the next 18 months

Asked what one thing they’d invest in over the next 18 months, panelists named time. Technology creates efficiency, but the question is how you reinvest the time it frees. One speaker put it simply: time is non-renewable, and if the tech means you can send another 10,000 emails, it’s worth asking whether that’s the best use of it. Their answer was to spend it on insight: sitting with colleagues to look at data and analytics, staying curious, and being in the right place at the right time, in something close to real time instead of a once-a-year cycle.

It’s the pharma version of something I keep coming back to: don’t stop at faster.

A short checklist

1. Replace one annual plan with a quarterly learning agenda that funds experiments.
2. Write down a test-and-learn policy: what can be tried without approval, and how results are shared.
3. List your stack components and mark which could be swapped within a quarter.
4. Identify one capability gap where a partner beats building.
5. Decide how you’ll reinvest the time AI saves, before it disappears into more output.

Building a test-and-learn AI program? Get in touch.