Stop using time-based trials

I've helped teams go from 10% to 90% retention. One of the first thing I do is go through their trials strategy. Let me guess one of your top support tickets:

"Can you extend our trial?"

I get it, time-based trials are easy to implement, clear to the user, and ROI is timed box.

But your leaving money on the table, here's why:

❌ Users either wait until the end of the trial to pay (money on the table) or ask for an extension.

❌ You force the user to get value in the time frame you think they should (different teams procure products differently).

❌ You end of charging them for just access to your product, unsure how much value they actually got, and can't price them up to the plan that fits their needs.

There is a better way: usage-based trials.

Let's dive in

The average B2B freemium product retains only 19% of its free sign-ups in M1. At BuildBetter, we retain over 60% of free users, and over 90% paid retention. How?

Before BuildBetter.ai, I consulted around the bay in various leadership product roles; almost always, the first thing I focused on was onboarding. I helped teams go from 10% M1 retention to 90% M1 retention. I saw the friction of time-based trials and teams unwilling to even entertain usage/value-based trials, often due to their potential disruption of KPIs. As a user, I hated them, and as a product leader, I vowed never to do time-based trials.

Why is our retention 3x better than the average?

There are many factors that go into retention, like the focus on day 1 activation, but one of the biggest blockers for getting higher M1 retention is time based trials.

Users sign up, 7-14 days go by, then the user no longer can use your product, even if they would have otherwise converted—and worst of all they probably went and signed up for a competitor and then converted because they already understood the value from using your product.

First, let's start off with a quick history lesson. At the end, I give you a basic formula you can do on your own product to see how much money you're leaving on the table.

Time-based trials.

Time-based trials are the default in SaaS today; effectively, a user signs up, and typically, 14 days later, they reach out to support asking, "Hey, I wasn't able to really use the product yet. Can you extend my trial?"

Sound familiar?

When you think about it, it's kinda silly. Why do we have time-based trials?

I actually couldn't find a concrete answer; the best guess I could find was:

• They're easy to implement, and before the internet was a thing, time-based trials could exist locally without many workarounds or server validation.
• It was easy for users to understand. Users often had very few programs they used, so a new one either worked or didn't work in their workflow fairly fast.
• There was a time when time = money, back when AOL charged you for minutes. Getting time was like getting free money.

Side note: I love how AOL had TWO time-based trials, both the amount of hours for free and the time you had to use those hours. They were a real believer in time-based trials.

But do ANY of those three scenarios above apply to most SaaS companies today? Sure, they're still easy to implement, but does that mean they work? Absolutely not.

Time-based trials today:

• Users either wait until the end of the trial to pay (money on the table) or ask for an extension.
• Force the user to get value in the time frame you think they should (different teams procure products differently). Sure, there's some urgency, but for larger purchases, that can be a huge turn-off.
• You end of charging them access to your product, unsure how much value they actually got, and can't price efficiency. That means they only had enough time to play around with half your features, and couldn't even see the value for your most expensive plan.

Yuck.

I'm not saying there's no case for time-based trials; just 80% of the time, it isn't applicable. So here's the alternative.

"Value-based" or usage trials

Welcome to the future. What is a usage-based or value-based trial? Why now? What changed?

First, value-based and usage-based are similar enough to be interchangeable in practice, but in theory, there is a key distinction.

Value-based: The user has reached some outcome or deliverable they find valuable in.

Usage-based: The user has done an action a set number of times (hopefully valuable).

In practice, value is often too vague, and so to measure the value, we often have to answer, "When has a user reached the value in our product?" which means finding some usage pattern.

In a perfect world, having a diverse set of different actions and outcomes being tracked verified by some qualitative evaluation of the user would be great... but in practice, we often just have to pick one or a few usage-based metrics that limit once they hit a universal governor where the user is then asked to pay.

Effectively, value-based or usage-base pricing, lets the user use your product until they've gotten value often measured by some key usage indicator.

Here are a few examples of companies that do usage-based trials:

• Slack (10,000 messages)
• AWS ($ of free compute) - If you're a startup you've seen this
• Zoom (45 Minutes Free Calls)

We now live in an era where we can track every action a user takes and know exactly what in our product is going to convert them.

Example

Here is a pho-company that right now does feature-based pricing in an industry that typically does time-based pricing but would be better off using usage-based pricing:

Let's say you were a video streaming platform like Netflix.

Netflix right now doesn't offer trials (at least publicly). This makes sense; everyone knows what Netflix is and has probably tried it. But, you're a new video platform, like Apple TV, and not everyone wants to pay right away.

Instead, the obvious first idea would be you may want to offer a trial, let's say 30 days.

The problem is, past user abuse, people would sign up and maybe not find anything worth watching in the month... maybe you're still ramping up your content. So, then extend the trial to 3 months like Apple TV does? Right?

Maybe, but extending the time doesn't mean customers will magically find value, or be charged when they get value.

What if they signed up, couldn't find anything interesting to watch, and a new show comes out 3 months and 1 day after they signed up? Well, you just lost a customer.

To address this, you have a brilliant idea to do a usage-based trial; this trial would only charge customers once they've watched a certain amount of shows or hours.

"Ah," your CRO says, "the problem is that some people may have watched shows they didn't like, while others watched a single show that changed their life. What then? So when do you pop up an upgrade?"

The ideal answer is to get ratings or look at some engagement metric per user, and dynamically charge based on some basic ML model that identifies when a user would be most willing to pay. So once they've thoroughly gotten into a binge-worthy show, present a pay pop-up to continue watching.

If you really want to get greedy, you would do it right before the final episode of some binge-worthy show.

But please don't share that idea with Netflix.

Realistically though, you can see how pure value-based pricing is kind of tough, when DO you decide to charge?

So let's look at usage-based. We'd want to pick some heuristic that generally applies to the "average" user, let's say 5 hours of watching shows, or 10 episodes, and then prompt a pay once a user has hit one or both of those.

You could get more detailed like if they watch a particularly high engagement show, or a particular show that only a person who can pay would watch, you can get crafty, but the same idea applies.

The Strategy

Find the thing in your product that you know once a customer does you have them, and then do everything in your power to get them to do that action. Not within 14 days, just as fast as possible. Once they've hit it, charge them.

This way of trials can:

• Align your team on TTV (time to value), making sure customers are focused on performing core actions in your product as fast as possible.
• Increase your conversion, as customers only see pricing pop-ups once they have seen the value of your product.
• Increase your ARPU. Customers may not be able to scale up the usage of your product in 7-14 days, but if you set the usage to the right metric, teams can get enough adoption of your product where when they do convert, they pick the higher priced plan or are now paying more overall on usage-based billing.

Not only that but your support team isn't hounded with "Can you extend my trial?" but is now having serious sales conversations about "you've used our product, you've seen the value, what more information do you need to move forward with a plan?"

Doing usage-based trials in conjunction with usage-based pricing is ideal for this conversation, but that's another conversation because the credits you may choose to give your customers are now tied to real $ value.

Trade Offs

It isn't for every type of business, below though, you can see how much more revenue you'd get by switching.

One trade-off we see at BuildBetter.ai is that the time to conversion can be longer for some customers (and shorter for others).

We're a call recording tool that turns otherwise untapped data—like call recordings with your team or customers—into your company's most valuable asset. Perfect for user research, sales, and team product calls.

Our usage based trial is hours recorded.

Some teams hit the hours recorded limit in the first few days, and others can take 2 months. But we know that whenever you do hit that limit of hours, you will pay.

You can imagine that some people have a user research project coming up a few weeks from now and want to explore a tool, or other teams have already recorded calls and want to upload them to explore right away.

We don't discriminate when you need to get value. We can convert both slow and fast customers without friction.

Now what? Should I do this?

Go to your metrics dashboard, look at your "magic moment." If you don't know what your magic moment is, you probably shouldn't be doing pricing optimization, but basically look at your paid customers or highly engaged customers, and see what they do that customers who churn don't (not an exact science).

Next, see how many customers have hit that magic moment but never converted, and how many customers never hit that magic moment because your time-based trial hit them with your time-based trial.

Make sure you're looking at cohorts not overall users.

This should be easy to see; you'll see a percentage drop off at the trial-end date per cohort. So after 14 days, retention goes from 50% to 10% etc. Take that delta minus the average day-to-day churn as the number of users you'd still have in your pipeline to convert at that 14 days date.

So let's say you lose 5% of users in a cohort everyday, and you had 50% retention at day 14. You'd have 45% of those users left on day 15.

Now look at your general conversion rate from the 14-day trial user conversion.

Let's say it's 5%. You can include the overall conversion, but it's best to use the people who only converted at the 14-day+ trial date. So let's say you convert 10% of free -> paid, but you only convert 5% of those users at the 14 day end-trial pay wall. Use that 5%.

Do a regression and see how many more customers or revenue you would get with usage-based pricing based on the conversion rate of the users now retained past that drop-off minus the average day-to-day churn.

Now times that new amount of customers by the ARPU of your customers. That's how much revenue you're leaving on the table.

So 45% of the users times 5% conversion times your ARPU.

Ideally you'd kinda see what it would look like day-after-day, include day-to-day churn, but that depends how much accuracy and convincing you/team need.

One last thing, remember, value-based pricing doesn't have to be the usage of an action a user takes; it could be seated (first seat free), revenue generated from your tool, or leads responded, you get the idea. Be creative, have fun.

Notable Mention:

Feature-based trials are also something people do, but I don't think it's worth diving into them because they aren't (really) trials. There is no "end of the trial." It's just an "if you want X other feature, then pay."

You could consider Slack to have a combination of feature and usage-based pricing in that you can get triggered to pay if you want a specific feature like integrations.

It's worth a mention, but it's fairly rare for higher-priced software to do purely feature-based trials. I think if you do have a feature based trial, I would push you to consider even more heavily a value/usage-based trial because you're half way there.

Typically a feature-based trial is based on the features that provide enough value where people will hit a roadblock where you can trigger a conversion (effectively, once they use X they'll want to use Y—when you give a pig a pancake model); you can see this transition done well for more freemium prosumer products, that have a "try before you buy."

Don't get me wrong, service like Canva lets you use their service for free forever but if you want access to "more" or "better" features you have to pay. It's not always a bad model, but Canva may have even more success if they let you use all premium features to a limit, v.s. having to pay to see if they like a specific feature.

Source for avg. freemium retention: https://openviewpartners.com/blog/your-guide-to-product-led-growth-benchmarks/