Tuesday, September 10, 2019

The Three Rules of Freemium




At the SaaStr Europa conference in Paris a couple of weeks ago I sat down with Joaquim Lecha, the CEO of our portfolio company Typeform, to talk about “Freemium at Scale”. Founded and headquartered in Barcelona, the company launched a free version of its service seven years ago. During our conversation Joaquim revealed that this free service helps drive 180,000 monthly signups, and about 3% of those signups convert into paying users who are billed anywhere from €25 to €70 per month, depending on the plan they choose. In other words, Typeform is effectively leveraging a free version of its product to drive paid subscriptions at scale.

I have a bit of a love/hate relationship with the freemium model. Done right, freemium can lead to amazing success. One of the best examples is Dropbox, which Tomasz Tunguz called the “King of Freemium”. What makes the company unique, he argues, is how it transformed its free users into evangelists. “Unlike other SaaS companies, Dropbox spends more of its revenue on engineering than sales and marketing,” Tomasz wrote. “Typically, businesses spend twice as much on S&M.” In a piece I wrote when Dropbox went public last year, I pointed to the section of the company’s S1 that detailed how Dropbox drove sales of its enterprise solutions. Unlike most other enterprise software, which traditionally used to be chosen by the IT department, Dropbox is typically adopted by individual employees from various departments, who then lobby management into switching. As I noted in my piece, Dropbox was one of the early champions of the ‘consumerization of enterprise software’ movement, which was one of the strongest drivers of SaaS success in the last ten years.

But not every SaaS company can be a Dropbox or a Typeform. Done wrong, freemium can end up cannibalizing your paid user base while also draining your company’s precious engineering and customer support resources. So how do you know if launching a freemium product is the right move for your company?

Let’s discuss some of the pros and cons of the freemium model.

The downsides of freemium


I think many SaaS companies are too optimistic in thinking that they can just offer a free, pared-down version of their software and that this will result in a wave of user signups followed by increased revenue once those users make their way down the purchase funnel. But there are a number of factors you should consider first:

Added costs: Given that SaaS is an extremely high-gross-margin business, one might think that you can easily support free users. However, even if your gross margin from paying customers is 80% to 90% (i.e. your CoGS are only 10% to 20%), those costs can become very significant if you grow a large user base that doesn’t generate any revenue.

From engineering to hosting, the freemium model will require consistent upkeep that drains resources that would be otherwise devoted to your paying customers. And even though your freemium users won’t be paying a dime, they will still expect some level of customer service. In our discussion at SaaStr Europa, Joaquim revealed that, on average, 70% of Typeform’s support tickets come from free users and that the company spends $130,000 per month supporting them.

Now, Joaquim didn’t consider this too high a price to pay for the various benefits of having a free plan (more on that below), but for other companies, the upside/downside assessment may look different. If your business has lower-than-usual gross margins (e.g. because your SaaS solution includes a service component or because your product is particularly costly in terms of infrastructure), you should think extra hard about whether freemium is right for you.

Cannibalization of paying users: For any freemium model, the running assumption is that a certain percentage of non-paying users will eventually convert into paying customers. But what should also be considered is how usage will flow the other way. In other words, some people, who without a free plan would have become paying customers, will be fine with the free plan and won’t need the paid version.

An imbalance of product features: A freemium approach requires a delicate balance. Provide too many features for free and you risk cannibalizing your paid user base. Offer too few features and you eliminate the value proposition for users to sign up in the first place. Typeform walks this tightrope well, limiting its free surveys to 10 questions and 100 responses. Converting to the paid version grants the user unlimited questions and responses, as well as advanced features such as logic jumps and design personalization options. Not every SaaS company can strike that kind of balance.

Less focus on core users: An important factor to keep in mind is that having a freemium model will almost inevitably have a strong impact on your product roadmap. If you have a free plan, chances are that for every paying customer you’ll have 10–20 (or more) non-paying users. It’s hard to ignore the feedback of a group of users that represents 90–95% of your total user base. Listening to those users doesn’t have to be a bad thing, but it may. It all depends on how similar your non-paying users and their use cases are to your paying customers. In this interesting analysis about the “rise, fall, and future of Evernote” — once the poster child of freemium — Patrick Campbell and Hiten Shah conclude that “trying to appeal to everyone and not building the functionality that core customers want to use made Evernote’s product feel stagnant, which is definitely a tradeoff they should never have made.”

Widening the top of the funnel also makes it critical that you are excellent at identifying the best leads effectively and efficiently. If you don’t do that, your valuable signups might fall through the cracks in all the noise.

The benefits of freemium


Most of the above challenges can be overcome if your free plan leads to a much larger top of the funnel and if you can convert enough free users into paid. A freemium model will likely lead to a lower conversion rate, but that’s OK if it’s more-than-offset by the increase in signups. Assuming that you can keep conversions at a sustainable level, then the freemium model can have several benefits:

More active users: One of the biggest challenges that SaaS companies face is driving adoption of their product. And while some users can be enticed by a free trial period, there’s a subset of consumers who are just more likely to keep using the product if it’s free. Others won’t even sign up in the first place if there’s no free plan.

More evangelists and a positive impact on your brand: You shouldn’t measure a user’s worth solely on whether they eventually start paying you. The larger the number of active users, the greater the pool of potential evangelists who will promote your product to new users. In my discussion with Joaquim, he told me that users who were aware of Typeform’s brand prior to signing up were twice as likely to convert into paid users than those who came in via non-organic channels. Qwilr, another Point Nine portfolio company, has made the same observation.

Amplify virality: The strongest rationale for going freemium is having a product with a built-in viral loop (like Typeform or Dropbox). If you’d like to dig deeper into viral growth in SaaS, check out this great post by my colleague, Louis. One caveat I’d add is that you can’t take it for granted that your free users will have the same viral coefficient (or k-factor) as your paid users. In many cases, your average free user will be less active than your average paying user and will therefore lead to fewer referrals. It doesn’t have to be like that, though. In an ideal world your paywall is built in such a way that users have an unlimited ability to share (or do whatever it is that makes your product viral) and monetize something else.

More user feedback: In our talk, Joaquim pointed out another advantage of having a large user base: It allows Typeform to learn from a much larger number of people. That’s a very interesting aspect that I hadn’t thought about before. Keep in mind, though, that as mentioned further up, depending on your product and industry, feedback from free users may be more or less relevant.

Re-engaging trial users: Every SaaS company will have a certain subset of users who will sign up for a free trial of the paid product and will not convert into a paying user once the trial period ends. Introducing a freemium version allows you to re-engage these users with the possibility of converting them at a later date.

Making the decision


So now that we’ve looked at the potential upsides and risks, how can you decide whether launching a freemium product is the right choice for your company?

Ultimately, only an A/B test can answer this question. However, getting reliable results will take a lot of time, especially if you want to measure the impact on virality and if you have a viral cycle time of, say, six months. If you can’t wait that long (or if you’re not equipped to do a complete-funnel A/B test), here are my “Three Rules of Freemium”:

1) Does your paid plan have a gross margin of 80–90%?
If you have a lower gross margin — for example, because your product is not fully self-service, requires extensive customer support or is extremely costly in terms of tech infrastructure — freemium will probably not work for you.

2) Does your free plan attract the right audience?
If your free users are too different from your paying users, your free-to-paying conversion will be low — and you’ll risk developing your product for the wrong audience.

3) Is your product inherently viral?
If your answer is no, that doesn’t make it a complete no-go, but it does mean that it’s much less likely that freemium is right for you.

Wrapping up

In the end, freemium only makes sense if a certain percentage of your free users do one of three things: 1) Eventually convert to paid, 2) refer paying customers, or 3) provide the kind of valuable feedback that will improve your product. A freemium product that fails to achieve any of these effects will merely saddle you with extra costs and distract you from servicing your most important users. Not every company can be a Dropbox, but the good news is that not every SaaS company needs to adopt Dropbox’s freemium model to succeed.


This post was first published on Point Nine's Medium channel.

Sunday, April 07, 2019

Five years later: Five ways to build a $100 million SaaS business

Back in 2014, I wrote a post titled “Five ways to build a $100 million business”. If you haven’t seen it yet, the central idea of the article was to look at how many customers you need, for a given ARPA, to get to $100 million in annual revenue and what this might mean for your sales and marketing strategy. That post went kind of viral, which led us to create a follow-up piece, an infographic, a poster (which you can order here), and a PlayPlay video.

What did I learn from that experience? First, if you want to write a killer blog post, it helps if you take a difficult question and simplify it drastically, give it a catchy headline, and add pictures of cute animals. ;-) More importantly, though, using my five little animals as a simple framework has helped me challenge (and hopefully occasionally provide sound advice on) the scaling strategies of numerous startups over the last couple of years.

Based on my learnings in the last years, here is a slightly updated version of the chart:


(Click for a larger version)

Here’s what’s changed.

Bye-bye, dear little fly!


In my original post, I spoke about five ways to build a $100 million Internet company. The post wasn’t specifically about B2B SaaS companies, which is why two of the five ways dealt with customers with an ARPA of $10 and $100, respectively. I will keep the $100 ARPA customers (AKA mice) for now, but today it’s time to say goodbye to the $10 ARPA flies. Bye-bye, dear fly, you’ll always have a special place in my heart. Thank you for adorning so many slides and posters – maybe someone will resurrect you for a consumer-focused version of this framework.

Hail the whale!


Five years ago, I thought it would make sense to use a $100,000 ARPA elephant as the largest animal on the chart. The main reason is that I grew up as a consumer software and consumer Internet founder, and even though I had already spent about five years as a SaaS investor when I wrote the post, my experience was heavily skewed towards SMB SaaS. I did include two larger animals in a followup post, but meanwhile, I think that the $1,000,000 ARPA whale deserves a place as one of the five prime SaaS animals. While there are only a few SaaS companies with an ARPA of around $500,000 (not quite a whale yet, but definitely much larger than an elephant) across the entire customer base (Veeva, Workday, Demandware, Opower,...), there are quite a few SaaS companies with a significant whale customer segment. Most public SaaS companies, unfortunately, don’t report any data broken down by different customer segments, but it’s pretty safe to assume that Salesforce, Box, Zuora, and a number of other companies derive a significant portion of their revenue from whale customers.

Why y is now x and x is now y


This is a smaller, somewhat technical change. The original chart showed the number of customers on the x-axis and the ARPA on the y-axis. Since it’s more customary to use the x-axis for the independent variable (and as I think it makes more sense to think of ARPA as the dependent variable), I have switched the axes. I had made that change in the poster already and have now updated the chart here as well.

Some animals are more equal than others


One thing I’ve learned over time is that just because there are five ways to build a $100 million business, it doesn’t mean that those five ways aren’t equally promising. To quote the pigs in George Orwell’s “Animal Farm” (one of the very few books that we had to read at school that I liked): All animals are equal, but some animals are more equal than others.

Let’s take a step back. The key take-away of my “5 animals framework” is very simple: if you want to get to $100 million in annual revenue you’ll have to find customer acquisition channels that are highly scalable and profitable. Otherwise, you’ll never get to the number of customers that you need at your given ARPA. The problem is that most customer acquisition channels are either scalable orprofitable but not both at the same time (which is why early CAC/LTV metrics can be so misleading, more about that here). Now, it seems like in some ARPA regions it’s easier to scale profitably than in others. This is certainly what we’ve seen in our portfolio: Several of our SaaS portfolio companies successfully went upmarket – oftentimes from rabbits to deer or even elephants – when they saw that they would hit a growth ceiling in their existing segment. One of the underlying reasons is that in order to get very large, you have to get your churn rate close to zero (or better yet, achieve negative churn), which is usually not possible if you’re selling only to SMBs; the other reason is that outbound sales doesn’t work if your ARPA isn’t large enough, which sort of limits your addressable market to companies that are more or less pro-actively looking for a solution like yours.


If you take a look at this analysis by Sammy Abdullah of Blossom Street Ventures (a VC with the tagline “We’re the anti-VC”, by the way) you’ll see that although Sammy points out that most SaaS companies don’t have an ACV of $50k or more, the vast majority of the 61 publicly traded SaaS companies from his list are deer or elephant hunters. Only 15 of the companies from the list had an ACV of less than $5k at the time of their IPO, and for some of them, the average is misleading because a large part of their revenue comes from customers with a much higher ACV (e.g. Zendesk, Box).

A recent analysis of private SaaS companies by Nathan Latka suggested the same. In his analysis, Nathan looked at 369 companies at around $1M in ARR. 186 of those (50%) are focused on rabbits or smaller animals (he really used those animal analogies). Of the 20 companies with $100M in ARR that Nathan looked at, on the other hand, only 6 (30%) are focused on low ACVs. If you are a successful rabbit-hunting SaaS company and you think you can keep growing in your current segment, these numbers shouldn’t discourage you, though. First, especially the numbers from private companies need to be taken with a grain of salt. And second, even if there is a statistically significant clustering of mid-market and enterprise SaaS companies at the $100M ARR mark, the data also shows that it’s possible to get there with a low ARPA!

[Update March 13, 2020: Here's a webinar that I did about the topic a few days ago.]