
I may live in a marketing-measurement bubble, but if you’re a brand — especially a consumer brand — you’ve likely felt the itch: “‘Maybe we should think more seriously about incrementality.”
The good news? You don’t have to overhaul everything at once. Here’s how to start.
Step 1: Recognize attribution’s limits
This isn’t just about being anti-attribution. Attribution (GA4, MTA, platform reports, etc.) has historically been the default source of truth for many teams. However, the first real step is recognizing what attribution is and what it isn’t.
Just because something gets credit doesn’t mean it caused a sale. And just because something doesn’t get credit doesn’t mean it had no impact.
Your organization may be overly reliant on attribution if:
- Marketing’s attributed ROAS always seems to increase, while total revenue and profit don’t.
- You’ve added a bunch of new channels for diversification’s sake, yet the ones above goal limp along at tiny budgets and the ones that don’t hit goal are paused, never to return.
- When you spend more, your blended CAC inflates. When you spend less, your blended CAC decreases, but your new customer volume remains stable.
This all points to the same thing: Your marketing spend (and team) is optimizing for credit, not growth.
Dig deeper: What your attribution model isn’t telling you
Step 2: Test in the easiest, highest-impact place
Don’t overhaul your entire measurement strategy tomorrow. Start with a clear hypothesis and a simple test. The ideal place to start is branded search.
Most people intuitively understand that branded search is not very incremental. It often cannibalizes organic traffic, and performance looks amazing in-platform. This makes it easy to get buy-in and provides a clear test structure.
Turn it off and watch what happens.
- Does organic search rise by just as much as brand search declines?
- Does revenue hold steady, or do you see a real drop in sales?
From that test, you can start building organizational muscle around what incrementality means in practice. You may also discover that this same effect is happening in other tactics — especially those that mix branded search or retargeting into their targeting.
Dig deeper: How attribution masks what’s driving growth
Step 3: Share results and win buy-in across teams
You can’t run a test, see a result and then carry on like nothing happened. If branded search ROAS in-platform is 25, but your test shows it’s 2.5, and you break even at a ROAS of 4, you need to communicate that and what it means for where dollars should be spent.
Ideally, there is a culture of curiosity and openness to pivot how things are done. Without that, this step is tough. Many teams bury results that challenge the status quo because they fear it will make them look bad. This is often why many incrementality test results get put into a deck that rarely makes it out of a marketing team meeting.
But ideally, results spark questions across the organization:
- How should results affect budget planning?
- Which channels are underfunded because of attributed ROAS and should now be tested for actual impact?
- In a period when we are more focused on growth, what should our budget look like?
- What about when we need to focus on profitability?
These last two questions, in particular, can forge a genuine partnership between marketing and finance.
Step 4: Turn insights into smarter budget decisions
Once you’ve run that first test and communicated the results, the next move is critical: change behavior. Measurement tools exist to help you make better decisions. They do not create value in and of themselves.
Use measurement data to confidently make bets and then measure the outcomes, knowing they are bets. Informed bets, but bets nonetheless. Use business-level metrics as your goalpost.
If branded search has little value but has historically taken up 20% of your budget, cut that spend and monitor your blended marketing efficiency ratio (MER) and total revenue. Ideally, revenue will hold steady while MER improves — confirming you’ve eliminated a costly, low-value tactic.
Now make a bet. Put that budget into a top-performing channel to see if there are marginal returns to be gained. Do you see your revenue increase or a spike in MER without meaningfully impacting revenue?
Relying on blended metrics might be enough for smaller brands. At larger brands, marketing may account for too small a share of total revenue to see quick results. In that case, design geo-tests to get a clearer picture of impact, then track blended metrics over a longer period to gauge overall success.
Dig deeper: Marketing results don’t add. They multiply and synergize.
Avoid these common pitfalls
Test and learn is good, but more importantly, you must test, learn, evangelize and act. The biggest mistakes are:
- Failing to bring others in the organization along for the incrementality journey.
- Failing to change behavior based on what’s learned.
Other common pitfalls include:
- You realize attribution is flawed, then overcorrect by throwing money at tools, vendors and dashboards you’re not ready for. Instead, start simple. Run one test. Use blended metrics.
- You panic when you see a range of possible outcomes (confidence interval) because you’re used to a single number. Understand what each end of the range means for your business. It will guide whether your next move is a slight optimization or a bigger pivot.
- Channel owners/teams get defensive when a channel gets tested or results contradict attribution. This happens when team culture is not aligned with the broader business priorities and marketing works in a silo. As a marketing leader, evangelize the risks of attribution and make it clear that marketing is aligned with finance and the company’s objectives.
In closing
Incrementality is the missing link between marketing activity and business outcomes. The best teams don’t just ask, “What did we get credit for?” They ask, “What moved the needle?”
More than an occasional KPI, incrementality is the backbone of how they measure success, plan budgets and optimize for growth.
The post 4 steps to kickstart incrementality without overcomplicating it appeared first on MarTech.
Go to Source
Author: Tom Leonard