A testing roadmap prioritises which channels to test, in what order, and how to connect each test's results to your budget decisions. Without a roadmap, tests happen reactively and the results rarely compound into a coherent view of your marketing effectiveness.
Most teams run their first incrementality test because of a specific trigger: a channel is underperforming, a new agency is questioning the media mix, or leadership is asking for proof of ROI. That is a legitimate starting point. But a one-off test produces one answer. A roadmap produces a measurement system.
Step 1: audit your current measurement gaps
Start by listing every channel you spend on and the confidence you have in its attributed performance. For each channel, ask: do we have evidence that this channel is causally responsible for the conversions it claims, or are we relying entirely on platform attribution?
Channels where you rely heavily on platform attribution with no experimental validation are your highest-priority testing candidates. These are also likely to be the channels where attribution is most inflated.
Step 2: prioritise by spend and uncertainty
- Start with your highest-spend channels. A 10 percent improvement in budget allocation for a channel taking 40 percent of spend moves the needle far more than the same improvement for a channel taking 5 percent.
- Next, prioritise channels with the widest gap between attributed ROAS and your intuition about true performance.
- Then, consider channels where your current MMM or attribution model has the weakest statistical basis.
Step 3: design tests that answer a decision
Every test on your roadmap should be linked to a specific decision. Before designing the test, write down: "Based on the result of this test, we will [decision A] if lift is above X and [decision B] if lift is below Y." If you cannot complete that sentence, the test is not ready to run.
Tests that answer questions nobody has yet decided to act on are a waste of time and budget. Measurement should serve decisions, not the other way around.
Aim for one major incrementality test per quarter per channel. Running too many tests simultaneously makes it hard to isolate channel effects and can contaminate results across tests if any channel suppression overlaps.
Step 4: connect results to your MMM
Plan how each test result will feed into your marketing mix model. If you run an MMM, schedule calibration updates after each completed test. If you do not run an MMM, use the experimental results directly as your channel effectiveness benchmarks for budget planning.
Over time, a well-executed roadmap produces a set of experimental benchmarks across your channels that form the evidence base for your annual budget allocation.
Step 5: communicate results and iterate
After each test, share the results with stakeholders in plain language. Avoid statistical jargon. Focus on the business question the test answered and the decision it informs. Document what you learned and use it to refine the next test in the roadmap.
A testing roadmap is not a one-time plan. It evolves as your media mix changes, as new channels emerge, and as your conversion volumes change. Review and update it at least quarterly.
How do I get internal buy-in for an incrementality testing roadmap?
Start with a single high-profile test on a channel that leadership is already questioning. A clear, credible result from one well-run test is more persuasive than a long-term roadmap presented in the abstract. Use the first result to build the case for a systematic programme.
How many tests can I realistically run per year?
For most marketing teams, four to six robust incrementality tests per year is a realistic target. Each test requires planning, execution, analysis, and stakeholder communication. Trying to run too many tests simultaneously creates quality problems and risks contamination between tests.
Should I include brand campaigns in my testing roadmap?
Yes, eventually. Brand campaigns are often the hardest to justify to finance because their effects are long-term and diffuse. An incrementality test on brand advertising can provide concrete evidence of its value, even if the lift on short-term conversions is modest. Longer test durations and brand awareness metrics are needed for these tests.