The current jumble of opinions on effective SaaS pricing, no matter how out of the ordinary, all spring from a default — dare we say, sacred — model. Subscriptions.
It’s THE wheel founders shouldn’t bother reinventing.
I think if you’re not willing to question existing systems, then you’re not in the right job, honestly. As a founder, your whole job is to question systems. So everything, in my opinion, is up for debate, even if it’s been done that way for 30 years.
Through the hard-won lens of their many — “painful” — business model iterations, an incredible understanding of their evolving customer segments, and off the back of sustaining a stellar (35% MoM) growth rate for 12 straight months, Rachel lays out:
— Why the assumed binary of PMF is impossible without this third element and why the subscription model wasn’t the right fit for Wethos and their customers
— How they trialled their way to a freemium hypothesis (and packaging narratives)
— The (segment) tension between subscription-based and transaction-based models, how Shopify solved for it from the opposite end, and why the future will be all about mixed models
— Finding predictability (that most-sought virtue of subscriptions) in a different SaaS model
— The constant battle all pricing iterations must reckon with
Some background. Before starting Wethos we had been running our own agency business, servicing clients for a couple of years. That business grew really fast. We had $1.4m in revenue in barely 18 months.
During those months we started to realize how hard it was to scale that business up. Which led us to develop our own software. Then we began wondering how we could release the software to help freelancers basically self-organize into teams so that we are not the central entity that’s kind of putting these curated teams together for our clients.
In 2020, when the pandemic hit, a lot of the market’s tailwinds and trends were moving in our direction. So we took the chance on winding down our agency business.
I believe that product-market fit isn’t a binary, black-and-white thing.
It’s a trifecta, instead.
Which gets completed only when you tie in and iterate as much on the third, equally decisive element — the business model. And the question a lot of people have on business models is: how are we going to make money?
But the question I have is: How are we going to provide value to our users and help them realize that value at the point of monetization?
To expand this line of inquiry:
- Where and at what moment are we going to get somebody to an aha moment?
- What is that aha moment actually worth to that person?
- How does that tie into their overall ability to succeed on the platform?
- And who are we actually going after, because that also has to map together, right? (folks you initially target as customers might be very different from those who can actually afford to pay)
We’ve done a tonne of price testing. Assessing different pricing models. And what we found was, for us, at least, the subscription model wasn’t going to work.
Because it raised too many questions that created too much friction. Our users were, of course, accustomed to the classic, Slack-like, per-seat model.
But being small studios and not full-fledged agencies, they scale their businesses up and down based on the types of projects that come across their desk.
Thus the typical per-seat model made no sense.
They were asking us questions like: ‘Am I going to have to pay per seat for the collaborators that I invite?’ ‘Are they going to have to pay a monthly subscription on their own?’
And we realized that we’re actually putting friction right in the place where we could be unlocking growth. If our customers aren’t inviting others to the platform because they’re afraid of whatever monthly subscription they’re going to incur, then we’re fucking up basically.
Unlike Slack, there was no ‘land and expand’ strategy to grow inside of a company. Each invited individual freelancer was a potential customer for us.
Then it’s a really tricky balance between targeting and knowing who your customer is. And then who is finding you and saying, ‘I’m your customer.’
At first the subscription model made more sense for our initial target audience, which was the high-earning independent segment, about 3 million freelancers in the US who earn over 100k a year.
Because of the pricing data we were publishing, we were getting an influx of demand from newer and greener freelancers; all the way to now, where 40% of our users are side hustlers. They still have full-time jobs. Paying a monthly subscription to run a side business just didn’t seem right.
That’s when we started to shift our focus over to trying out a transaction-based model.
And the key question around transaction based models that every founder has to answer for themselves is: how do you get a margin on payments?
Because if you integrate with a Stripe, or any third party like that, you’re not gonna get any margin. Stripe’s charging 2.9% already, which means we’d have to charge upwards of 4% to make any money on that. So while 40% of our users are side hustlers, 35% of our users still make over 100k a year and they’re not going to give up that much money. They’re just not going to want to.
That’s totally insane.
So you see, the same product with a different audience changes the economics. For greener freelancers the higher transaction-based fees make more sense, for the high-earners a subscription trade-off would be a better deal.
Thus a major hurdle for us was how are we going to be able to do this with enough of a margin to keep payment processing fees low and to give the software away for free (more on this in the following section), so that we could unlock distribution for the widest possible range of freelancers.
General rule of thumb is to narrow your audience instead of trying to service them all. But our product was resonating with everyone. And whether we were ready for it or not, the growth was here and we needed to find a way to meet the demand.
Luckily, my co-founder, Claire, our COO, is just amazing. And we met Unit, our banking service provider, in December of 2020. And Claire basically worked with them to come up with a model where we could embed banking and payments and debit cards into the platform.
Which already made sense for us because the other issue with this payment processing side of our platform was that we were trying to facilitate a complex payments flow.
A one-to-one-to-many flow.
Which meant, a client is able to pay an invoice into a user’s account. And then that user needs to be able to disperse that money at their discretion to their teammates.
Long story short of it is, we found a great partner in Unit that enabled us to offer 1% on ACH of getting an invoice paid. And we had a really great deal with them, where we have a really wide margin underneath.
As ACH banking is free. But it’s a pain to go into your bank account and pull that information out and actually do all that. We make 1.5% on debit card interchange as well. And then we make a little bit of the interest on money in all the users accounts.
Once we were able to find the right partner to unlock that model, that’s when there was a bunch of discussion around how to monetize the volume itself.
Do you do it on payments, transactions, transfers in and out? Or peer-to-peer? A lot of different ways to slice it. We have essentially boiled it down to a hyper-growth model.
So peer-to-peer payments on our platform are free and instant. You have to have a bank account to set up on the platform, but other than that it’s free to pay others and to get paid.
We only make money when they start making money and getting invoices paid from their clients. Once we were able to give that stuff away for free, give them access to all that pricing data, and then include the peer-to-peer payments.
That really started to unlock virality.
That’s when you have that trifecta: the market’s demand is there, the product experience makes sense and the business model is tied closely to user behvaior and value. So it’s easy to understand the value and start paying for something when you actually experience that value
Rather than psychologically trying to trick someone into converting.
We launched that free model in February of 2021. And we acquired 150 users that month. Right now, we’re acquiring 150 users a day. It has completely transformed the business. This past month alone, we’ve acquired roughly 5,000 users.
So it’s been really fast and furious.
None of it, of course, happened overnight.
We were actually making some money off the subscription model we went live with.
Before we adopted freemium, what led us there was that our 14-day free trial was way too short for our user’s businesses. Their sales cycles are much longer.
We were looking a lot at Shopify as an example. Because that’s the best comparison we had in the beginning. Shopify entrepreneurs sell physical products and our entrepreneurs sell services.
We saw that Shopify had a 14-day trial, but quickly found out that setting up an e-commerce business is a very different value chain in general.
For a freelancer, sales cycles can take weeks, rather than selling potentially every day in the case of an e-commerce business. You might send out 5-10 proposals a month, then half of those close and turn into projects, and then those projects turn into maybe three-five invoices each, per month.
It’s just a very different thing.
We realized that we were seeing a ton of growth, top of the funnel, but the retention within that 14-day period was really struggling.
So the first thing we did was to make the trial period longer. We tested 30-day trials, then I think we tested 60 as well. Which is hard, because there’s a lag in results.
You need some data and information to understand if a change is working or not. When you extend the trial periods for all the users who signed up on January 1st, and then others who signed up on January 15th, then we had to wait another 30/60 days for those cohorts to start converting.
Thus measuring by cohort can get really challenging. As honestly, at a startup you can either choose time or money. You cannot have both. It’s a race.
We only had so much time before we ran out of money or until we could unlock enough growth to raise the funding that we needed to really kick things off.
Then eventually we began asking how we can offer the platform basically for free?
Given the feedback we were getting from early customers around the per-seat model, when we launched our Network features in February 2021, we thought that would be the perfect time to launch freemium.
Because we were basically saying: Now as a freelancer, you can join Wethos and you can not only scope your own projects, but you can get discovered by other studios.
Most agencies today are three freelancers under a trench coat. Just freelancers hiring other freelancers.
We wanted to wrap our message around that insight. Which was: At Wethos, our vision is to make running a studio both easier and accessible. And on that accessibility track, we’re giving the platform away for free.
A better narrative. Because the other thing that’s really challenging about price testing is that you’re failing in front of people. Without a resonant narrative how do you explain four pricing changes over the course of two months?
We owed it to ourselves to sit down and really think about it. How do you not just launch the feature as soon as it’s ready, and instead craft that into a narrative that is going to help with growth, and we were able to do that with the Network features.
So we launched a big database where you could create profiles, you could check out our scope library, check out our services, you could set all these things up for free.
That’s when things really start to take off.
But yeah, there was a lot of testing and learning in-between.
Pain, I would say, in-between.
This is another thing I looked at Shopify for. And what’s interesting in looking at Shopify’s history is that Shopify started the opposite way that we did.
They started with transaction-based fees.
Then they found out that the people who were really successful at running Shopify stores hated that. Because they were making a lot of money and didn’t want to trade off those fees.
The people who were not so good at it, although they were signing up a tonne, they were bringing too much of that classic SMB churn, so Shopify was not making money off those people.
Which is when they introduced a subscription model for the folks who were making more money on Shopify. The trade-off there amongst a lot of other features and what not from Shopify, was a lower payment processing fee.
For us, what we’re looking at now is trying to segment not based necessarily on a demographic, or psychographic, but more on just brackets of earnings.
And saying ‘okay, once you’ve hit this threshold of earning over 100k a year, even paying us 1% On that is kind of bullshit.’
If you’re paying us $1,000 a year for that, we could charge, whatever, 50 bucks a month, and you could be saving money by trading off and switching over to the subscription model.
And I honestly think that the world is probably going to move in the direction of a lot of mixed models. We’ll have more features for a more premium tier. Plus there are other things that we can do in a more traditional sense.
In terms of what we’ve heard from the users and how they interact with the platform, the biggest trade-off is saving money on payment fees, because, again, the other thing is volatility and your income.
If you’re earning over 100k a year, you get a pretty steady predictable income month over month. And so it’s easier to add a fixed cost like 100 bucks a month to that.
If you’re still just getting started or things are volatile or unpredictable, then having that subscription to your on your card is like death.
You’re nervous you won’t be able to pay it. That’s kind of how we have been thinking about evolving the pricing structures for different segments going forward.
I understand that the monthly recurring revenue, first of all, is an amazing thing to have. Because you know, already, that next month, unless you literally add no net new customers, you’re still going to bring income in and then potentially continue growing.
For me, the thing that’s awesome about a SaaS model is not just predictability. But it is the stability of recurring and only having to focus on net new users added each month.
The key assumption a lot of the time is that once you’re on the subscription, you are going to keep coming back. The hard thing about more transaction-based models is the inherent risk of people not returning.
For us, though, I don’t find it to be less predictable than the SaaS model that we had. The reason for that is because a transactional funnel just has — although the conversions are different — a similar dynamic.
So when I look at our funnel, before, with the subscription model, it was: How many people have started their free trials? And then how many people convert into paying customers after they start those free trials?
Let’s say if a free trial is over, and you only converted 5% of those people.
Well, that’s it for the other 95%.
So just like any other SaaS product, we needed more time to keep engaging customers and keep getting them to explore new features.
Eventually what we discerned from that is: What are the actions that users are taking on the platform that have the highest propensity to lead them to getting an invoice paid? And how do we optimize for that, basically?
If you’re a pure neobank, onboarding a new customer, your process is going to be more about activating that user through using your debit card or transferring money into your account. For example, when you’re signing up for Robinhood, the first thing they have you do is connect an account and transfer money in, to start trading.
So the hard part was that we are sort of a bank, but we don’t market ourselves as a better bank for freelancers, because I personally just don’t think anybody cares for, or thinks about that.
What we market ourselves as is software to run your studio. Again, the narrative really matters here. You don’t want to have this weird narrative violation where you’re like, come in, start scoping, and then ‘please transfer money into your business account.’
The best path for us we found was: Come in > Create a scope of work > Turn that scope to an invoice > Send yourself that test invoice.
And now we know that a higher percentage of people who are exposed to that and onboarding are going to come back and actually become a real transacting user.
We did an experiment last year where we proposed to users that if they would send us a test invoice for $10, we would pay it into their account. So that they could see what the whole experience was like.
Because what we found out was that what they were the most worried about was the experience their client was going to have with paying the invoice, not necessarily how easy it was for themselves to use it. Which is fascinating.
Basically for us, I know, predictably, right now that we’re growing 35% month over month, each month, and January is a year straight of month over month growth, which is really hard to do at the transactional level.
I know now that if I pour this many users up top, and this many users are then onboarded, and exposed to these many features, then X percent are going to go on to transacting.
We also now know that after a user gets one invoice paid, the majority of them come back and start running their business on the platform.
That, to me, can be as stable as a SaaS model and sets us up for additional upselling or cross-selling in the future
Once you have that hook, and then you see the retention. There’s no reason for them to churn after that, because they’re fully embedded now in the system.
Clearly similar dynamics as SaaS.
But everything in my mind comes down to: how do you sell this in a way that actually makes sense to people instead of just kind of copying and pasting what other people are doing?
I think this is a continual thing that everybody struggles with.
Which is building the features that people are asking for/want versus building things that are going to help unlock the business.
Ideally, most of the time those two things are aligned. But there are certain constraints around resources. For instance, our full-time staff even right now is only 14 people.
So at this moment in time, I only have five engineers.
When I say to the team, ‘okay, we need to extend the 14-day trial.’ Or, ‘we need to shut the subscriptions down completely, untangle Stripe from the whole situation.’
And then get this transactional business model up and running through an API (luckily, those APIs were really well documented) of setting up bank accounts with engineers who had no experience in fintech, but are very amazing people and can Google their way through anything, it can still get really frustrating.
And although, I think we do a decent job of getting everybody on the same page about why this is happening, it’s still a race against time, and changing your revenue model will look more like a pivot than a product launch to investors.
So we go from ‘okay, we were going to prioritize this feature that people were asking for to no, actually, we need to put engineering resources towards implementing freemium.’
The same constraint appeared over the last six months. All of the playbooks on earth essentially say, focus on the first time user experience first, that is what unlocks growth.
While pushing out these pricing iterations, we ignored stuff that we knew might help retention, or that people were asking for. It’s very painful and hard to do that when you have thousands of existing users who want certain things.
We spent months testing and refining that new user onboarding experience. And now we’re sort of turning a corner and starting to focus a lot more around retention.
You can only do so much at once.
So the constraint is always going to be focus, and then resources.
A constant battle…
— Flodesk’s co-founder, Martha Bitar, on deciding against an industry-standard free plan
— Nira’s co-founder, Hiten Shah, on the one thing he’d do differently with pricing
— Geckoboard’s founder, Paul Joyce, ponders whether pricing is more alchemy than science