CRO is Killing Your Reach
· Articles

A founder told me something on a call this week that I've been thinking about since.
They sell a subscription product. They'd been working hard on improving the economics of first-time orders, pushing people toward a two pack instead of a single, removing the one time purchase option from paid funnels. The financials looked better. Payback period shortened. LTV held up.
Onto a winner right?
Then, slowly, CPMr started climbing. New customer acquisition cost went up. Every week felt harder than the last.
Raising the barrier to entry shrinks the audience Meta can find
This is the mechanism. When you increase the minimum commitment required to buy: higher AOV bundle, subscription-only, you don't just change the order economics. You change who is willing to convert.
Meta optimises toward people who are likely to complete a purchase event. The more expensive or involved that purchase event becomes, the smaller the pool of people who will complete it. The algorithm finds that smaller pool, serves them, and your CPM per person reached goes up.
The conversion rate falls. AOV rises. On paper it looks like a reasonable trade. But reach is quietly compressing. You're paying more to reach fewer people. And CPMr rises not because of anything you did in the ad account, but because of what you did on the website.
The problem compounds before you see it
I shared an example from another brand on the same call. A single-product supplement company that had made a series of CRO changes in January, pushing toward higher-bundle entry points, restructuring the offer. The financials looked fine. New customer AOV was up. The media buyer was happy.
By January the following year, they were looking at an audience that had been quietly getting smaller for twelve months. CPMr had drifted up. New customer reach was down. The retention was propping up the numbers but new acquisition had been slowly suffocating.
Nobody noticed because nothing broke suddenly. It just got harder, week by week, until you couldn't ignore it.
Filtering subscription renewals from conversion events made it worse
The same brand had tried to solve a separate problem, noisy attribution from subscription renewals showing up in campaign data, by filtering renewal events out server-side. The logic was clean. We don't want Meta to see renewals and think the campaign drove them.
The result was the opposite of what they expected. CAC went up.
The reason: Meta's bid in the auction is partly based on estimated action rate, which is its confidence that a given creative or campaign will drive a conversion. When you flood the account with subscription renewal events, Meta becomes more confident in the account. That confidence translates to a better bid. Better bid, better delivery, lower CPMs.
Filtering the renewals out felt like cleaner measurement. What it actually did was reduce the signal Meta needed to compete effectively in the auction.
Attribution looked cleaner. The account performed worse.
New visitor percentage is the early warning you're missing
On a separate call the same week, different brand. Sunglasses, three product lines, too many campaigns live for the spend level. Spend up 8% year-on-year. CPMr up 27%.
The first thing I checked was new visitor percentage. If it was low, I'd suspect poor exclusions. If it was healthy, the problem was structural, too many campaigns, too much overlap, self-competition in auction.
New visitor percentage was 77 to 80%. Fine. The exclusions weren't the issue. The fragmentation was.
New visitor percentage tells you whether Meta is going out and finding new people or cycling through warm ones. It's the most direct proxy for whether CPMr is rising because of something structural or something in the account setup. Most brands don't track it.
The CRO team and the media buying team need to be in the same conversation
This is the deeper point from both calls.
CRO decisions are almost always evaluated in isolation. Does the new bundle structure improve AOV? Does removing the single pack increase subscription rate? These are legitimate questions with real answers. But they don't account for what happens to the ad account when the answer is yes.
Every change to the offer changes who is willing to convert. Every change to who is willing to convert changes who Meta can find. Every change to who Meta can find changes CPMr.
The most useful question to ask after any CRO change is: did spend grow faster or slower than reach? If spend grew and reach stayed flat or fell, CPMr has gone up. Something in the conversion event or the account structure is narrowing the audience. It might be the CRO change. It might be fragmentation. It might be both.
The metric is available. Most people aren't looking at it.
Read more about CPMr:https://taikundigital.com/phil-kiel/cpmr-calculation
Stay up to date or book a call:https://taikundigital.com/linkbios/phil