You’re staring at a dashboard. It’s glowing. The "Total Revenue" line is trending upward like a hiker on a mission, and your heart rate settles into a steady, satisfied rhythm. But then a voice in the back of your head—probably your CFO’s voice—whispers a single, annoying question: "How much of that would have happened anyway?"
That’s the soul of an incremental.
In the messy world of growth marketing and business analytics, an incremental represents the portion of a result—usually sales, clicks, or sign-ups—that was caused specifically by a particular action and wouldn't have occurred otherwise. It’s the difference between "I sold 100 shirts" and "I sold 20 shirts because of that specific Instagram ad, while the other 80 people were coming anyway."
Most people get this wrong because they confuse correlation with causation. Just because someone saw your ad and then bought your product doesn’t mean the ad made them do it. They might have been your mom. Or a loyal customer who buys every Tuesday regardless of what's on their feed.
The Brutal Reality of Baseline Sales
Let's be real for a second. If you turned off every single ad tomorrow, your business wouldn't immediately drop to zero. Unless you started the company five minutes ago, you have a baseline.
🔗 Read more: Convert Chinese Yuan to US Dollars: What Most People Get Wrong
A baseline is the gravity of your business. It’s the organic momentum you’ve built through word-of-mouth, SEO, and people just remembering you exist. To find out what an incremental is, you have to subtract that baseline from your total results.
Imagine you run a pizza shop. On a normal Tuesday, you sell 50 pies. You decide to hand out flyers on a street corner, and that day you sell 65 pies. Your incremental lift is 15 pizzas. It isn't 65. If you calculate your Return on Ad Spend (ROAS) based on 65 pizzas, you’re lying to yourself. You’re taking credit for the 50 people who were already hungry and walking toward your door.
Why incrementality is terrifying for some marketers
Honestly, it’s a bit of a career killer if you aren’t careful. Marketing agencies love to report on "Last-Click" attribution. It makes them look like geniuses. If they show a "10x ROAS," everyone pops champagne. But if you run an incrementality test and realize that 80% of those sales were "cannibalized"—meaning you paid for clicks from people who were already searching for your brand name—the 10x suddenly looks like a 2x.
It’s the difference between being a "Value Adder" and a "Credit Taker."
How We Actually Measure This Stuff
You can't just guess. Well, you can, but you'll probably go broke or get fired. Real experts use a few specific methods to pin down the truth.
🔗 Read more: John Cooper Han Capital: Why That Michael Jordan House Buy Wasn’t Just a Publicity Stunt
Randomized Controlled Trials (RCTs) are the gold standard. It’s basically a science experiment for your bank account. You split your audience into two groups:
- The Test Group (they see the ad).
- The Control Group (they see nothing, or a "ghost ad" about a charity).
If the Test Group buys at a 5% rate and the Control Group buys at a 3% rate, your incremental lift is 2%. That 2% is the only thing your marketing budget actually achieved. Everything else was just "background noise."
The "Intentional Dark Period"
Sometimes, companies get brave. They do what’s called a "holdout." They pick a specific city—let’s say, Des Moines—and they turn off all Facebook ads there for a month while keeping them on everywhere else.
If sales in Des Moines stay exactly the same as the rest of the country, guess what? Your Facebook ads weren't being incremental. They were just "tagging" people who were already buying. This happened famously with eBay years ago. They realized they were spending millions on brand-term search ads (buying the word "eBay"). When they stopped, their traffic didn't drop. People just clicked the organic link right below the ad instead. They were paying for what they already owned.
The Math Behind the Madness
If you want to get technical, the formula for incremental lift is $Lift = \frac{Test Group Conversion - Control Group Conversion}{Control Group Conversion}$.
This gives you a percentage. It tells you the relative increase. If you want the actual $Incremental Revenue$, you take the total revenue from the exposed group and multiply it by that lift percentage. It sounds simple, but the data is usually "noisy." People switch devices. They clear cookies. They see an ad on their phone and buy on a laptop. This "cross-device" behavior is the reason why measuring an incremental is becoming harder every year, especially with privacy changes like Apple’s iOS 14.5 update.
Where Most People Trip Up
There is a huge difference between Incremental Reach and Incremental Revenue.
👉 See also: Target Employees and Shoppers Losing Faith in the Company: Why the Red Bullseye is Fading
- Reach: Finding new people who haven't seen your brand.
- Revenue: Getting people (new or old) to spend money they wouldn't have spent otherwise.
You could have massive incremental reach—showing your ad to millions of new people—but if none of them buy, your incremental revenue is zero. Conversely, you could show an ad to a past customer (zero incremental reach) but offer them a "limited time" discount that triggers a purchase they wouldn't have made until next month. That is incremental revenue.
The "Halo Effect"
We also have to talk about the "Halo." This is when an ad for Product A causes someone to buy Product B. Or when a TV ad makes people search for you on Google. If you only look at the direct "click" on the TV ad (which doesn't exist), you miss the incremental value.
Economics experts often look at Marketing Mix Modeling (MMM) to solve this. It’s a statistical way of looking at all your spending over years to see how different channels interact. It’s less "real-time" than a dashboard, but it’s often much closer to the truth.
Actionable Steps for Finding Your Truth
If you're worried you're overpaying for "organic" sales disguised as "paid" sales, here is how you fix it.
- Run a Brand Search Holdout. Turn off ads for your own company name for two weeks. Watch your organic traffic. If organic goes up to compensate for the lost ad clicks, stop spending so much on your own name.
- Use Geo-Testing. Instead of testing "audiences," test locations. It's much harder for data privacy rules to mess with a geographic test. Compare "Ad On" regions vs "Ad Off" regions.
- Survey Your Customers. Ask "How did you hear about us?" It’s old school, but "Post-Purchase Attribution" surveys often reveal that the "last click" on a Google ad wasn't the reason they bought; it was a podcast they heard three weeks ago.
- Stop Worshipping ROAS. Shift your focus to iROAS (Incremental Return on Ad Spend). Only count the revenue that the tests prove was "new" money.
The reality is that an incremental is elusive. You'll never have a 100% perfect number. But by acknowledging that your baseline exists, you're already ahead of 90% of the people burning money on digital platforms. Start treating your marketing like a lab experiment rather than a slot machine. Demand to see the "lift," not just the "total." Once you start measuring what actually moves the needle, you can stop paying for the people who were already going to love you anyway.