If I had written this story last year, I would have gotten burned at the stake for my hot take. But I missed the mark on being a contrarian, seeing as today, that take is getting colder as more people agree.
Removing our free tier, replacing all of our CTAs with "Book a demo," and “forcing” people to talk to sales seems criminal, and makes a PLG enthusiast's eyes bleed, but it’s been one of the most successful experiments we’ve run.
The most ironic thing about this is that our product, Command AI, helps other companies be more product-led. We have prospects coming to us with exactly this need in mind. We sell tools that make your product easier to use, but can’t make ours sell itself? Not really, no.
For one, Command AI isn’t a bottom-up SaaS product, meaning that someone on your team can’t really try it for themselves and then invite others to join (think Figma, Notion, Airtable).
Well, in theory, they could, but it just doesn’t make a lot of sense. You don’t use Command AI individually to solve a personal problem at work, you use Command AI as a team (whether that’s the product, marketing, or customer success team), to create many non-annoying experiences for your users, run experiments to see which are most effective, and iterate over time as your product evolves. On top of that, implementing us requires a good amount of organizational trust – we’re going to show up in front of your users, after all.
That said, we are free tier people. Whenever we try products that require us to go through a demo before getting access, we sometimes groan and grumble. Ugh if the product is so good, why not just show it to me?
In the battle between vibes and business physics, we decided to run with the latter. So we ran a “no free tier” experiment. Buckle up and I’ll tell you about the results.
Our sales-led experiment
What were the problems we were trying to solve with it?
- We weren’t seeing as much exploration as we expected on our free tier. It was mostly really small companies that had no intention to graduate. (This makes a lot of sense. There’s always more urgency when you know $$ is on the line.)
- As a result, not many companies converted from free to paid. You could argue we didn’t give this enough time (time for our smaller customers to grow into companies that could benefit from our paid users). But ain’t nobody got time for that (to justify a PLG strategy at least).
The hypothesis
Removing the free tier will improve our trial conversion rate.
The results
It’s not often that you get very clear results from running an experiment. We did. Not to say there weren’t any negative reactions from removing our free tier, but we felt that for the time being, the positives outweighed the negatives.
The good
- We got to talk and learn from our prospects more. This allowed us to properly qualify product-solution fit. It also meant we got the chance to show off our awesome customer support.
Note: From my POV, if you’re looking to take something away from this experiment, this might be it. Too many companies (especially pre-PMF) get infatuated by the promise of PLG they’ve seen work so well with other products. What they forget is that those products (Loom, Figma, and the like) are mature companies that have achieved scale. They know their customers and they know the mechanisms that work to convert them. If your company isn’t there, then don’t pretend it is. There are many other things you can replicate, but this is not one of them.
If you’re early stage, talk to your customers and get to know them. I wouldn’t even try to have a free tier in the beginning – the risk of using the free tier to justify not talking to users is too great
Unless your product is the textbook definition of bottom-up, you have a clear plan for converting users from free to paid (and it works), and you’re continuously talking to customers, I wouldn’t consider opening up a free tier. And even then, I would still try the sales-led approach to see how much new stuff I could learn from prospects.
- Deal size - without much of a substantive change in top-of-funnel, opportunity size increased by 34%. This was driven, we think, by requiring prospects to follow our sales process more clearly. Because they couldn’t get into the product fast, we were able to get more senior people involved in the conversation, which is the key to deal size for us.
- Opportunity conversion – conversion increased by 17%. The specter of losing access to the product puts some (healthy! we think) pressure on evaluators to make up their minds during the trial. Of course, there are plenty of good reasons for extending trials, and we often do. But a free tier is like a backstop infinite trial (albeit with limited functionality). Removing that backstop created urgency. We created a lot more urgency for those on a trial.
The bad
- Some complained, which is to be expected and 100% valid.
We want every product, whether they are making revenue or not, to be able to use Command AI. But we have to prioritize. Right now, we’re seeing super exciting growth on our more expensive enterprise tier for bigger companies. We’re trying to stay ruthlessly focused on growing that. Until we feel like (a) we’re more in “cruise control” on that side of the business or (b) a free tier could actually contribute to paid growth via conversion (which it didn’t before) we probably won’t start up the free tier again.
In the meantime, I decided to talk to some founders and operators who have done the same thing so I can find out two things: (1) why they decided to do it and (2) is there anything they’ve learned through it that other companies could learn from?
Equals
Around the same time we started our experiment, we saw Equals publish this story on their experience removing the free tier. I’m not going to lie, It was pretty validating. Coming from a place where all you hear is “PLG is the future,” seeing someone else admit it wasn’t working for them made us feel even stronger in our decision.
Bobby Pinero, the CEO and co-founder of Equals, did a fantastic job telling their story in the article, but I was curious about the nitty-gritty details that didn’t make the cut.
“When we first launched Equals, we were all ‘book a demo / get on a call with us.’ We made you pay up front and we learned a lot from that. Once we opened up the product for free, we stopped learning. It was like we were flying blind. We didn’t know what people were doing, what they were trying to do, and what issues they ran into,” Bobby shares.
Another interesting thing that comes up is how the number of use cases for Equals grew as they stopped talking to prospects. In theory, this should mean a good thing, right? But more often than not, the product doesn’t (and shouldn’t) actually fit all those use cases. “Ever since we’ve gone back to a sales-led motion, we just say no to people [who don’t fit our use case] and don’t waste their time knowing they won’t be successful using our product.”
Trying to be everything for everyone is how many startups die.
Knowing what I know about our experience with PLG, I ask Bobby whether he thinks they’ll ever go back to having a free tier.
“Equals is a pretty horizontal tool. There are a lot of different ways you can get into the product so there are very specific ways in which we’d have to build onboarding for those different routes. The benefit of us doing a sales-led approach and talking to people is learning all those routes. Over time, I think there’s a path (when we’ve done it enough times and have had enough reps) to be able to open up some self-serve channels. But it’s going to be a while.”
When I asked him about any frameworks he recommends to those trying to figure out which sales motion works for their company, he shared the following questions you’d have to ask yourself.
- Do you have a massive potential user base?
- Do you have a really short time to value? Minimal setup: minutes, not hours.
- Is your product foundational for the end user?
- Do you have a very low incremental cost to serve each customer?
- Do free users contribute to your growth model? Are there viral loops that free users contribute to or accelerate? Are there network effects?
Pylon
Pylon is a pretty young company and they’ve always been sales-led. For context, Pylon helps support teams manage customers in Slack, Teams, email, etc.
I chatted with Marty Kausas, Pylon’s co-founder and CEO, to find out what made them stay away from PLG. His position is very clear.
“We actually have a pretty strong opinion on this: early-stage companies should never be PLG. Maybe with some exceptions, but so much of what you're doing is about learning and you lose a lot of that if you don't talk to people, you don't hear what they want or watch them set up the product. We think that’s a mistake for early-stage, even for small deals.”
Another point that I haven’t actually seen come up much in other conversations I’ve had is what needs to happen before you go PLG.
“Making your product self-serve is a lot of work. Thinking about the UX and the entire onboarding flow is hard. Is it the right thing for us to focus on making the product self-serve or should we focus on other things? For now, it’s definitely other things.”
At the risk of repeating myself, I do very much agree with Marty’s stance on early-stage startups: PLG is oftentimes an excuse not to talk to users.
Airbase
The next company I talked to, while still a startup, is a lot more mature. Founded in 2017, Airbase, the spend management platform, raised $250M+ and has always been sales-led.
I sat down with Darcie Lamond, who leads Content and Brand at Airbase, to ask her more about this decision.
“Our product is just not a great fit for PLG. Airbase is a very complex product. You have to implement it, it has to be integrated into all of your workflows. You have to get all your vendors and employees onto our system. That’s not something you just do without help or planning.”
Something I was curious about with Airbase is that a lot of their competitors have tried a product-led approach. How and why do they do it?
“It’s a question of how you want to build a company. There are products in our space where PLG works because they’re leading with just one use case, like corporate cards. But that was never really who and what we wanted to be. We’re building a far more complex and comprehensive product and I think taking that approach for the sake of being PLG would’ve been a distraction,” she shares.
I appreciate when companies stay true to their core beliefs and don’t jump on the latest trend because everyone in the space is doing so. Airbase doesn’t seem to shy away from the truth: companies need to generate revenue. Not something we used to hear a lot in the old days (and by old I meanthe ZIRP era, you know, like, 3 years ago), believe it or not.
“This kind of freemium model, it always seemed a little dated to me. You end up with a lot of people that have the product, but who aren't using it because they haven’t paid for it, so why would they care? We’re building sophisticated software and we never wanted to be a free product. We want people to want to pay us because they see the value in the product.”
Dots.dev
Dots is another YC company, this time from the S21 batch, which offers a developer-friendly API that lets businesses pay their contractors through just a few lines of code.
Sahil, one of the founders of Dots, tells me they initially started as PLG, but quickly realized it didn’t align with their goals: "We found product-led growth to be limiting, particularly in how it structured user access and feature availability. Our product had a broad range of features, but customers generally needed only a subset. Adopting a flexible pricing model based on individual needs helped us secure deals that a rigid PLG model would have hindered."
Something he shared felt very similar to something we’ve experienced at Command AI: “Our decision to be sales-led also stems from the necessity to get buy-in from senior stakeholders within companies. This is usually the case with more complex products.” With Dots being a fintech, this makes a lot of sense. “In certain industries, personal relationships and direct sales efforts are super important for securing significant deals,” he adds as another reason for why they made the switch.
Conclusion
Hope this take was useful. PLG isn’t for everyone, and the sooner you realize it, the quicker you’ll learn. If you want to chat more or want to challenge my take on why the types of companies we talked about should be sales-led (at least for some time), hit me up.
Until then, if you’re in the market for a product that makes your product easier to use, and don’t hate talking to humans, book a demo with Command AI.
