Thought I would have a little fun with the headline ;)
I was drafting up my Google Ads playbook last week, but I had a big "aha moment": before we can talk about Google Ads, we MUST talk about incrementality.
(P.S. If you want to get my Google Ads playbook next week you should sign up for a membership now).
The goal of marketing is to influence consumer behavior. You want to create purchases and spend that would not have happened in absence of marketing. That’s incrementality.
Before we get into it I want to thank this week's sponsor: Morphology Consulting. Morphology is a digital commerce consultancy that helps brands unlock incremental growth without spending more on ads.
A lot of digital marketing “best practices” around measurement muddy the waters of incrementality. You can make a lot of money as an agency, software vendor or advertising network by taking credit for conversions that would have happened with or without you.
When I share playbooks for different ad channels my goal is to help you maximize incremental spend. That’s the only way you’ll grow your business.
Here is the context you’ll need to understand why I’m making these recommendations. I’m doing this because your current agency partners/media buyers/etc. might push back on my recs if they’re not incrementality-minded.
Incrementality Simplified
Let’s pretend that you’re running a pizza restaurant located at the end of a long, winding road. You’re the only business in this location; no one is coming down that road unless they want pizza.
You’d like to bring in more customers, so you put out an open call for marketers to help you drum up some business. Your open call produces three proposals.
Marketer #1 pitches billboards placed along the two lane highway that folks must exit to reach your pizza restaurant.
Marketer #2 tells you that he will place flyers on cars in the parking lots of other pizza restaurants.
Marketer #3 proposes that you install an inflatable wavy arm man 100 yards down the road to your restaurant, and he’ll be holding a pizza pie in one arm.
Which proposal would you choose? What marketing activity is going to bring in the most new business for the dollars you invest?
It seems obvious that proposal #3 isn’t going to do bupkis for your business. You’re targeting people who are already driving down the road. Once they turn off the highway, it’s almost certain that they’re coming to see you. The “marketing investment” doesn’t influence their decision.
But what if you tallied up the transactions of everyone who saw each marketing campaign and measured them that way? In that case, the wavy arm man would look like a slam dunk, because everyone who visits the store sees him on the way in.
Believe it or not, that’s how most marketing measurement works: you see the number of transactions placed by the folks who saw or clicked on the ad. This doesn’t tell you anything about your viewers’ intent or the true impact of your marketing.
Clickthrough and view through attribution do not measure incrementality.
***
Meta ads is one of the most popular eCommerce go to market strategies, but it's getting harder and harder to make it work each year.
Rising CPMs, competitor brands and consumers who are spoiled for choice make it harder for brands to turn a profit online. And the DTC "peanut gallery" only ever seems to have one solution: spend more on ads.
What if I told you there was a better way? Morphology Consulting specializes in helping brands grow without spending more on ads.
Every Morphology project starts with a comprehensive audit to identify high impact, low effort growth opportunities for your brand. On average, Morphology clients improve revenue performance by 600% YoY.
Click here to reach out and learn more about how Morphology Consulting can help your brand scale without spending more on ads.
***
How To Measure Incrementality
The best way to measure marketing incrementality is a holdout test:
You split your target population in half. One half of the population receives your marketing campaign (the test group). The other half does not (the holdout group). Other than the marketing campaign, both populations are treated exactly the same.
You measure the conversions and spend from the test and holdout groups. If the test group spends more, or converts at a higher rate, and that lift in sales/conversions is statistically significant, your marketing “worked”.
Unfortunately, running a perfect holdout test isn’t always possible for a few important reasons:
First: running a campaign against a large enough audience to achieve stat sig is cost prohibitive for many smaller brands.
Second: a perfect holdout test requires the marketer to hold all other variables constant except the test variable. This means you can only test one campaign at a time unless you’re working with a large audience/large budgets. Again, this isn’t optimal for smaller brands.
Third: the marketer also needs access to the behavior of the holdout group to calculate results. This isn’t possible if your target population is prospects (ie folks who have never purchased with you before).
There are workarounds for these issues. For example, in geo lift testing, you use lists of demographically similar ZIP codes to create test and control groups. Then you measure the conversion and spend from each pool of ZIP codes to determine the incrementality of the campaign.
Even with workarounds, you have to weigh the benefits of perfect knowledge against the opportunity cost of moving slowly. To maximize your growth, you have to use your gut to make some marketing decisions.
Viewing Ads Through The Lens Of Incrementality
Telling you to “use your gut” is pointless if you aren’t used to running marketing through the lens of incremental impact. So here is a quick guide to the relative incrementality of different popular marketing channels:
Google Search
“Branded search” aka running ads against searches for your brand’s name is just about the least incremental marketing there is.
Non-branded search and other Google Ads products can drive incremental demand. But setting up your campaigns to exclude branded searches is critical, and Google has made it harder to do this.
Meta Ads
Conversion campaigns that explicitly exclude existing customers are usually quite incremental unless your brand has built up a lot of awareness outside of Meta.
ASC campaigns don’t allow you to exclude customer lists explicitly; instead, you set an “existing customer targeting cap”. ASC campaigns tend to target more existing customers than “traditional” campaigns.
Explicitly targeting your own customers might drive incremental spend from that audience, but the upside is limited relative to customer acquisition and should be measured carefully.
Retargeting Display Ads
If you have a large customer base and diverse sources of traffic, retargeting can potentially drive some incremental sales from new and existing customers.
Retargeting became less effective post-iOS14. Perform holdout tests and do not lean on reporting provided by the vendors here.
Affiliate Marketing
Coupon sites are one of the least incremental channels out there. Why do you want to pay a commission on sales from highly motivated customers looking for discount codes?!? Do not let these sites participate in your affiliate program.
Other affiliates can drive incremental sales. Monitor the breakdown of new vs repeat customer sales from each affiliate channel. Some affiliates work better as a “Middle of Funnel” marketing strategy.
Email Marketing
Sending more emails is the best way to drive incremental spend from your email file. If you’ve maxed out this strategy, email is more about shifting demand into certain days/categories vs driving incremental spend.
Email is the easiest channel for running holdout tests, so use it liberally to test offers and other strategies you’re considering rolling out elsewhere.
Final Note: most ad products do their best to blend high intent and low intent traffic together to make their in-platform performance look as good as possible (so you’ll spend more money).
Other ad products and software base their entire business model on intercepting high intent conversions and taking credit for them. If you’re testing some kind of javascript-based website widget, always run a holdout test (i.e. install the widget on half of your site sessions and compare the conversion rate and revenue/session for both groups).
If You Made It All The Way Down Here...
I need your help to keep this newsletter as relevant as possible. Can you click on the ONE future topic you're most interested in:
A. Playbooks for non-Meta marketing channels (Google, affiliate, etc.)
B. Meta deep dives (creative diversification, advanced media buying, etc.)
C. Breaking into and scaling Wholesale
D. Merchandising for DTC (expanding your product assortment, assortment architecture etc.)
E. Optimizing your site (CRO, site merchandising, etc.)
The links all go to the same page, but your response will be recorded.
I'll probably cover everything here, but I want to know what you care about most. If it's something that isn't listed here, reply to this email and tell me what's on your mind.