4 Incredible ROI Projects for (Most) Subscription Products
2023 is a different year, your projects have to make money now and not 3 years down the road
If you run a product team with 5 engineers, 1 engineering manager, 1 designer, and 1 PM, a 2-week sprint probably costs you around $50k to run if everyone is in the US.
The projects that you ship need to pay multiples of that money back. Ideally quickly. If each sprint doesn’t produce at least $50k of revenue, you will eventually run out of money.
The big projects help, but so do the small projects with immediate impact.
These are all projects we launched at Codecademy that had a great effort-to-reward ratio. I have replicated them with consulting clients and seen the same results.
If you haven’t done these yet, you should. Even if they don’t work, you’re only out a few days of engineering time.
1. Re-writing your checkout page error messages to be clearer (1–3 days)
The fastest way of increasing conversion is to help unblock the highest intent buyers.
Helping someone buy what they already want is easier than convincing an uninterested person to buy.
The highest intent people, by far, are those who tried to pay you and hit an error.
You do it like this:
- Download a list of the checkout errors (both what the user sees and associated system errors) for the last 90 days
- Dump them into Google Sheets/Excel and sort by frequency
- Go through that list and re-write each one to ensure that users know:
- Exactly what went wrong
- Exactly how to fix it
- What to do now
- Create a mapping table from old message -> new message
- Give this to an engineer to update
A bad error message is: “There was a problem with your payment method.”
A good error message is: “Your CVV code didn’t match, please update it and try again.”
A bad error message is: “Your card was declined.”
A good error message is: “Your card had insufficient funds, please input another card and try again.”
Try to hit the top 80% of errors that users see. This should take ~1 day to download, sort, and re-write, and then it should take 1–2 days for an engineer to change and release.
Because checkout pages are the only place where 1% of conversion = 1% of new revenue, this immediately impacts the business. I have seen this add 1–5% to checkout page conversion, which means you’re increasing the company’s revenue by 1–5% in under a week.
The more people you have purchasing your product each week, the more effective this is.
2. Turning On (or Up) Your Revenue Recovery Features (1–2 days)
A huge part of recurring revenue business is getting ahead of payment issues. For B2C applications (not charged through an app store), it’s common for 5–15% of your monthly recurring payments to fail.
With the right recovery steps, you can win back 50–60% of these charges, which makes a huge difference.
If you can fix a payment issue, you make that money back immediately and get the future earnings from that user, which you would have lost.
Almost all payment processors (Stripe, PayPal, Adyen, etc) and subscription management tools (Reculy, Chargebee, etc) have this functionality.
However, it’s not always turned on by default, which is a potentially huge loss for you. In most payment/billing systems, these changes can be made without an engineer.
You want to make sure that you have the following set up:
1. Turn on automated retry on payment fails, ideally a dunning process
Automated retries are exactly what they sound like; you set up the number and frequency of attempts after a card fails.
A Dunning process is a machine learning-based retry system that considers various data points and tries to “guess” the right time to retry the card.
In a recent client project, I have seen their dunning be 2x as effective as automated retries in Stripe. This is basically free money and worth upgrading to their “Scale” tier to access it. I’d turn on the Dunning setting as high as they let you. As of writing this, that is 8 attempts in 30 days.
2. Turn on & optimize your payment failure notification emails
You’ve probably seen notifications from products that ask you to update your card after it was declined.
Some payment systems have this on already; others don’t.
I would suggest:
- Make sure these emails are on
- Turn up the frequency just slightly below the “obnoxious” level. I’d go with 7 emails in 30 days, with more being sent at the beginning of the period.
- Put some love into writing those emails so they don’t feel so robotic.
Ideally, the place you send them doesn’t force them to log back in to update their card.
Note: I wrote a longer guide on this topic here: https://www.subscriptionindex.com/blog/payment-processing-101
3. Setting All Purchase Points to Default to Annual Plans (1–2 Days)
Most of the best ROI projects sound so simple that you assume there should be more to it.
There typically isn’t. That’s what makes them great.
Here are the steps:
- Find all of the purchase points
- Set them to default to annual plans.
Ideally, the design of your page works in a way where the user can easily compare the savings they are getting.
We made this change at Codedex (here and here), effectively bumping up the annual plan %.
Quick note: If you are building a freemium product, remember that a lot of where your user makes the decision to buy will not be on your pricing page itself.
There are likely a lot of upgrade points deeper in the product that you’ll have to remember.
4. Turn Off Your Monthly Email Receipts (1 Day)
This tactic is half a step into the gray area for subscription tactics, but I would argue that it’s industry standard.
Monthly email receipts are just nudges for unhappy users to cancel the product.
Netflix, Spotify, and all the other major players don’t send them for a reason.
- Turning off your monthly email receipts
- Setting all your purchase points to default to annual plans
Just a note that I recommend you keep receipts on for annual plans (as they might be higher tickets) and also keep the “your free trial is about to convert to paid” emails on as well.
The potential brand damage by charging people a lot of money without telling them isn’t worth it.
So What Should You Do With This Information?
There are a handful of scenarios where these tactics don’t work (e.g — if you only sell annual plans, then that is already the default), but the rest are worth a shot.
Worst comes to worst, you lost 4 days of engineering time.
The best-case scenario is that you’re getting some quick wins that compound for years.
