7 Reasons Cost Per Enrollment Is the Only Ads Metric Built for Online Schools

We are in a very healthy spot in terms of the return on ad spend. We are also enjoying quite a high profit. But it’s very important to look under the hood, because here we are just looking at the total orders, and when we look at what those orders consist of, we see things that we didn’t see before, we see trends that we were not able to see before.

Cost per lead measures who raised a hand. Cost per enrollment measures who paid, and for a school, only one of those numbers should set your bids.

1. Cost per lead answers how many people raised a hand. Cost per enrollment answers how many people paid tuition, and only one of those pays the bills.

Profit per student is actually a great way to look at this, because our numbers and efforts are mostly around making people buy. And they do buy, but only one time. So considering that we already paid for them, it’s good to explore other channels. And I’m bringing up the sales team here because we can advertise to the students that already passed through the school. But we will pay twice for the people that we already paid for. I’m making this point that it may not be worth it to pay for the same person multiple times, but profit per student is a good measure for that.

That’s the same math behind cost per enrollment: a raised hand costs money before you find out whether it turns into a paying student.

2. Your real lead-to-enrollment rate is different for every program, every channel, and every season, so a single blended number hides more than it shows.

We were enjoying and looking at relatively similar volume of orders. Year over year, the total number of orders didn’t change much, and it’s hard to say that it moved substantially in any direction. But what happened inside of that sample is much more important. The flagship product, the portion of people buying the more expensive option, is dropping, and the portion of people buying less expensive things, like re-enrollment, is growing. So we see the total volume of orders at more or less the same place. But when you just look at the flagship product, the one that used to account for the biggest share of orders, there’s an obvious decline in the buying trend. This metric, this dynamic, directly supports and explains the numbers and trend that we’re seeing. So relatively similar volume of orders, but under the hood, the distribution of things that people are buying is changing very substantially.

The same split shows up across programs, channels, and even the time of year, which is why running multiple programs means pricing it per program instead of trusting one blended average.

3. Tell Google to optimize for “form submitted” and it will find you more of the cheapest inquiries, which are disproportionately the ones that never enroll.

Google is explicit that only your primary conversion action drives bidding: a secondary action is “for observation only… not for bidding, even if the goal they are included in is used for bidding” (support.google.com/google-ads/answer/11461796). Set “form submitted” as primary and that’s the only number Smart Bidding ever chases, whether or not the person behind it ever enrolls.

If we didn’t initiate the quarterly review, we would be under the impression that it’s actually quite good. We have a good number of orders, more or less the same as last year. The return on ad spend is very healthy, and the profit margin is in line, more or less, with the previous year. But it didn’t show the full picture.

4. Cost per enrollment sits in the metric tree; cost per lead is a ratio one step removed from it, so watching the ratio and expecting the tree to move is backwards.

It’s an important question, what metric we will be using as a North Star. We were using the orders, we were using the return on ad spend, we were using profit. But it’s important to have all of this in line. We were mostly looking at the graph and comparing year over year revenue. Again, what do we want to look at? This is an important question, because this will determine the lens through which we will be looking, and we’ll be strategizing and optimizing the campaigns.

Cost per enrollment sits inside that lens as an outcome, alongside the other numbers worth tracking. Cost per lead is a ratio calculated one step away from it, the same blind spot behind a stuck CPA.

5. You cannot manage growth and efficiency with the same number, and cost per enrollment is the one that actually tells you which you’re getting.

It would be worth looking at new accounts instead of orders, because at some point, if a lot of people just stop extending, we can eventually find ourselves in a situation where revenue drops very abruptly, and we may have quite a high efficiency, but the total volume of gross profit will suffer. We are now enjoying the orders that are happening from the students that already bought, and we are looking at the total number of orders. But since there are less people buying the new courses and we’re just seeing the recurring purchases of the previous courses, at some point there can be a cliff, there can be a very abrupt drop. So it’s important to see all of that and decide on the next step.

Efficiency can hold steady on paper while the number of new students underneath it quietly shrinks. A cohort-based bootcamp feels the same tension between enrollment deadlines, and cost per enrollment is the number that actually tells you which one you’re getting.

6. A California online school built its ads around enrollment, not leads, and tripled its student count on the way to 47% of the California market, up from around 13%.

We worked with a California-based online school that rebuilt its ads around enrollment instead of leads, tripled its student count, added multiple 7 figures in profit, and grew to 47% of the California market, up from around 13%. The fuller story is here.

7. The fix already exists inside admissions: the number of people who actually enrolled.

There’s this pie chart: the sum total of all orders. Some of that is online, where people engage with whatever they see and then make a purchase. But the orders by admin, the net new orders that are happening over the phone, seem to be lagging behind. This is the comparison with the previous year, and we were thinking about this number. This may not necessarily mean that something is off with sales in general. What this means is that the online process became more smooth. It’s important to note that the orders by admin, there are a lot of orders just to extend the program, like extension.

Google describes the exact same situation: “sometimes, an ad doesn’t lead directly to an online sale, but instead starts a customer down a path that ultimately leads to a sale in the offline world, such as at your office or over the phone” (support.google.com/google-ads/answer/2998031). Import that admissions number as an offline conversion, and Smart Bidding finally sees the same enrollment your sales team already counts.

Where cost per enrollment starts

This is the setup we build before we touch a single bid: PPC management built around enrollment, not raw leads.

Send us the account and we’ll tell you if the number you’re bidding on is the number that actually matters.