Top 10 Learnings about Free Trials with Tomasz Tunguz

Ready to go. Awesome. So my name is Tom Tungus. I'm a managing director at Redpoint. I've been there for about 10 years. I write a blog@tomtungus.com that focuses a lot on early stage startups and uses a lot of data in order to illuminate different topics. Redpoint is a venture firm that's been around for about 20 years. We invest from the very earliest stages to the latest stages of software and consumer companies and we're based in California. Over our 20 years we've been lucky enough to work with over 15 different unicorns and some of the most notable companies in the last generation of software, including companies like Stripe that just raised at 22 and a half billion, HashiCorp and Snowflake, which are two of the fastest growing software companies in history and some of the next generation businesses like Looker and others. A lot of these businesses use a common marketing technique in order to grow really quickly and that marketing technique is free trial. Being on the board of some of these companies, we're often asked the question what is the most effective way to structure a free trial? And, and for those of you who read the blog, you know that I love data. And so what we did is we went out and we created a data set in order to answer this question. Late in October of last year, we sent out a survey to understand how the best marketers and marketers across all different kinds of SaaS use free trial in order to acquire customers the most effective in the most effective ways. That survey had 24 questions and we were overwhelmed by the response. We had almost 600 different companies respond of all shapes and sizes and I'll get into the distributions of that for a second. We took that data set over the last three months and it was an enormous amount of work. We wrote over 1000 lines of code in R and we performed lots of different statistical tests to make sure that the result of the analysis I'm going to share with you actually stand to statistical significance to give you a sense of the distribution by arrow. Of the 600 or so respondents, about half of the about 2/3 of them are 5 million in AR or under. But we had five or six different companies that were publicly traded well north of 100 million who are using free trial in order to continue to grow their businesses. In terms of ACV, they span the gamut everywhere from 5k all the way to 150k plus and they were broken down roughly a third, a third, a third by targeting the SMB, the mid market and the enterprise. In addition, we believe that this data is representative because if you look at the different kinds of buyers from the respondents, you see it spans everything from operations and marketing to sales and engineering, even down to legal. So we really believe this data set is comprehensive and provides at least some sense of what's really going on in the wild when use of free trial There are two goals of this analysis. The first goal is to share with you benchmarks. We believe benchmarks are really useful to help you build your business because they provide good goal posts for financial planning and for goal setting. And the second goal is to help you have conversations within your business about what the right mechanisms are to test for retrial. The way that we structure this presentation is top 10 learnings. So let's jump right in Contract length. How long should a contract be in order to maximize conversion rate? If we look at the distribution by respondents here on the on the Y axis we broke it down by four different categories. Usage based how many API calls. In other words, you can use as many different API calls as you want. A multi year contract, a month to month contract, an annual contract and on the x axis we broke it out by ACV from 0 to 5 all the way to 150k. What we see is the way to read this chart is 80% of respondents in the 15-50K ACV use an annual contract. And what you'll notice is that looking at the heat map and what we've got is in the mid market the vast majority of respondents actually use an annual contract. In the SMB, 2/3 use month to month and in the Enterprise you've got 2/3 using multi year. What does this mean? It means that for most companies an annual contract tied to a free trial is the best way of going about it. And it turns out if you run the correlation of the conversion rate by free trial, there's actually no statistical difference. So even if you're the SMB and even if the month to month contract is the most popular, you really should be using an annual contract because the conversion rate from your free trial won't change. So conclusion number one observation number one from this data set is stick with annual contracts, they're industry standard and there's no change in conversion rate. Our second topic, benchmarks around retention logo retention logo retention is starting last year how many customers did I have? And from that cohort this year how many have I retained? If we look at the distribution across the respondents, we see 2/3 of the respondents actually retained better than 80% of their logos. The top third maintained 90% plus. So in building your models for your business, you really should be striving for 90% plus logo retention rate. Oops, here's another slide. So if we break down the so one of the key questions that comes up a lot is if I'm an SMB company or if I'm a mid market company, or if I'm an enterprise company, should my logo retention rates change? And the answer is yes. What we notice is that in the enterprise it's much more common to retain 90 to 100%. 45% of respondents who target the enterprise maintain 90% of their logos. If you look in the mid market, it's pretty similar. But if you look at the distribution ESMB, it's actually quite different. You have almost a fifth of respondents failing to retain 60% or more of their logos. That's a very difficult way of building a business. So the conclusion number two is you really should be striving for 90% logo retention rates. If we move on to net dollar retention, net dollar retention, also called negative net churn, also called account expansion. If I take a cohort of customers who are paying me $1,000 last year, how much are they paying me this year, then it's net of churn, how many customers have I lost and the revenue associated with it and the customers within that cohort that have actually expanded or grown their spend with me? If we look at NDR again, distribution by percentage of companies, we see that the vast majority, about 60% of companies are somewhere between 80 to 120% net dollar retention. One of the things that we say in venture is I was just in a board meeting where we had this conversation. You know, the very best companies are really expanding their accounts 120% plus or more. The top quartiles right there. If you look 120% and above, you've got about 8% that are doing 140% net expansion or better. That's top decile. Again, let's look at the same chart that we looked at in the previous buy logos. And if we look at the enterprise, what we find is it's much easier or it's much more common within the enterprise to actually attain higher levels of net dollar retention than it is in the SMB. In the SMB, in order to get to top quartile, you have to go all the way down to 100% net dollar retention. Where in order to get top quartile in the enterprise, you're looking at 120 to 140% net retention. Why this surprised me, this was one of the results where we've invested in SMB companies like Expensify that have phenomenal net dollar retention growth rate, net dollar retention numbers. And so I thought, you know, broadly across the industry that you would have lots of different SMBs that had really great NDRs. But when we stepped back and thought about it and talked to some of our companies, the reality is if you're selling to an enterprise, you have way more seats to sell. There are way more departments to sell to. There's just a lot more people and a lot more business that you can get. And so it's far easier, as we saw in the logo retention to retain those customers and also to expand them. So it another observation was you really should be setting a net dollar retention target of 100% to 140%. That will put you within the top either third or decile or quartile depending on the category Trial structure okay, so there are four ways to limit a free trial. The first way is through features, compliance features, security features, Slack uses search. The second is through seats. Expensify did this in the early days. Your first two seats are free. Once you want to get to the third seat, it's time to pay. Time is the third one. 30 day free trial, 14 day free trial and then the last is usage. How many API calls do I get free with Twilio before I need to pay? We broke this out by ACV and what we find is across every single acv, time is the most frequent and most commonly used limiter to trial. The second most common actually depends changes as a function of the acv. So what we find is usage is very common in small ACVs and as you go up, seats in the enterprise are much more common as the delimiter. The question is what is the impact on conversion rate? Well, it turns out time and usage trials actually convert twice as effectively as features and seats. And this is statistically significant. Right? So if you are structuring a free trial, you really ought to be considering a time bound free trial or a usage based free trial rather than a feature based free trial or a seat based free trial. That's observation number four. Let's talk about trial length. If time is one of the most effective limiters of free trial, how long should your free trial be? Should it be 7 days, 14 days, 30 days? 90 days? 120 days? Some people have unlimited free trials. So it turns out that 14 day free trials are by far the most common. I don't know why, and I don't know why only 3% of people use 21 day free trials. It's really just a week of difference. It turns out it doesn't matter. The conversion rates across all those lengths are statistically identical. By the way, I'll be sharing these slides on tomthinkers.com so feel free to take photos. But they'll all be published later today. So what we find is if the conversion rates across the free trial links are exactly the same, why lengthen it? You should just shorten your free trial. And the theory there is, if you're engaging with the buyer who's using a free trial, all of the momentum, all of the desire to use the product is really right there at the very beginning. And the longer the time they're engaged with the product yet, the longer the time they have to decide not to use it. Time kills deals is another aphorism we use in venture. And so if you've got the point of maximum intent at the beginning, use a shorter free trial. Salespeople, should you use salespeople? Should you hire salespeople in order to should you hire salespeople in order to call your leads? Well, it turns out 75% of the respondents actually do this. Obviously, this is more expensive, right? If you're hiring a salesperson, an inside sales rep in the Bay Area is probably gonna cost you 120k. A field rep is probably gonna cost you $250 to $300k, maybe more, depending on their quota. But this is a very common practice and there's a reason it turns out that you're almost 4x in the conversion rate. So if we look at unassisted conversion rate across the entire respondent set, about 4% of incoming leads through free trial actually convert to paid. But if you have a salesperson call them, that goes up to 15%. Again, a really statistically significant result. The question as a business is at what point, at what point, at what price point does it actually make sense to hire a salesperson in order to make the cost of customer acquisition scale? And really that comes back to your price point. So there's another blog post that's on the blog that talks about how you really need to be kind of a 10 to a 15k price point and make sure that you're paying your inside sales reps an appropriate amount in order to make sure that this works. But if you're above that price point, there's no question that you should be Using salespeople in order to call those qualified leads. And what we find is that this observation actually is true across all different price points. So unassisted conversion rate by price point. Unassisted conversion is. The median is 7% and 0 to 5%. A 0 to 5K ACV, it's 3% and a 5 to 15K ECV. You'll notice that as the price points go up, unassisted conversion rates go to zero. This means that as people want to spend more money, they want to talk to somebody, they want to be reassured, they want to have some sense of trust in the company that they're buying from. And what you'll also notice is that conversion rates actually dip through the through increasing price points. But in every single bucket, assisted actually meaningfully doubles or triples or even quadruples the effectiveness of the free trial. So hire salespeople to close leads, moving on to conversion. So there is this number of 4%. Again, it's kind of part of industry jargon, which is that's the target. If I have an unassisted. If I have somebody who comes to use a free trial, what fraction of the free trial do I actually convert? Well, I've been batting around this number for as long as I've been in venture, which is about a 4% target. Here's the actual data. The 25th percentile, the bottom 25th percentile converts 1%. And this is unassisted. That means not being touched by a salesperson. The 50th percentile is 4% and the 75th percentile is 12%. And if you look at this chart through a distribution, it actually follows a power law, which means that, you know, most, most of the time, it's basically flat. And at the very end, you have a handful of companies that have really terrific product market fit and are able to convert these free trials at much more effective rates. But what that really means, and what that means for you is you really should be aiming at 4% or a little bit better in terms of unassisted conversion rate. If we look at the same analysis for assisted conversion. So the conversion with salespeople by quartile, what we find is the 25th percentile is 6.8%, the 50th percentile is 15.5, and the 75th percentile is 30%. The 75th percentile takes one third of inbound leads and converts them into paying customers. That's pretty exceptional. You don't see that very often. So what does that mean? Conclusion number eight, you really should be shooting for at least 15% assisted conversion rate on your free trials. Activity qualification. A lot of companies with their free trials, what they do is they look at product metrics in order to understand which of the prospective leads are engaging with their product and they use that to score leads. How many times does this person use this particular feature? How many times have they invited their colleagues in order to use the product? How many searches have they executed? All kinds of different activity metrics. As you're looking to qualify your leads and figure out which of the ones your salespeople should be calling, is this a good thing to be doing? So again, if we go back to the data, on the Y axis you have use activity to score leads. Yes and no. And on the X axis you have all the different price points. What we see is what you'd expect in the SMB when you're selling to small price points. Activity scoring is actually very good. Oh, sorry. It's basically the same. So if you use activity scoring to determine leads, your conversion rate is. It's actually not very good. My apologies. If you use activity scoring in order to discriminate leads, there's actually no real difference. So there's no value in using it. At the SMB. If you go into the mid market, what you find is that activity scoring actually meaningfully decreases. It actually meaningfully de. I'll say that again, it actually meaningfully decreases conversion rates. And in the enterprise, it does it to an even greater effect. So what does this mean? Using the data about how people are engaging with your product during the free trial and actually is bad in the sense that you are throwing out leads that otherwise could be really good. Why is this? Well, in the SMB, people are buying products the way that consumers buy products. But as you go more and more into the enterprise, the enterprise buyer is not going to be using the product necessarily. The buyer and the user might be different. It might be it versus the line of business owner. And so consequently, if you're using activity metrics in order to score your leads, you're looking at people and you're expecting them to use the product in a particular way in order to qualify their demand. When in fact what you're doing is actually qualifying out what could be a really great buyer. So really question activity scoring in the enterprise. Take a look at that. That's really important. Okay, the last one. Payment. Only 12 companies require payment info at the start of a trial. This was really low. I was really surprised by this figure. And so what this means is most of the free Trial forms today. What they do is they optimize on the total top of the funnel. They say we want to get the maximum number of leads, we want to get as much possible data about each individual lead. And so we don't require a credit card to be to be entered. But what we do find is if you do require payment, you actually more than double the conversion rate. What we weren't able to get a sense of is how does this impact the top of the funnel. In other words, what is the trade? If I implement payment as a requirement in the funnel, do I lose half of the top of the funnel or do I lose a smaller fraction? And so every company is going to be slightly different, but it's definitely worth testing because there's a meaningful increase in conversion. Awesome. So that's the 10, those were the 10 observations that we were able to pull from this survey. There's a lot more data here. We're going to be publishing the slides@tomtengoose.com There are a lot more insights that we'd like to share with you and we will be in the future. But these were the ones that we wanted to cover today. So just to summarize for you the top 10 observations from this 600 person survey. The first is stick with annual contracts. They're by far the most common. They're an industry practice and particularly in the mid market, they're very common. There's no real meaningful impact to conversion rate for your contract length and so having your customers stick around longer is a net positive. The second is as you build your financial plan for this year, for next year, you really should be targeting 90% plus logo retention across most segments. Except for the SMB, 90% retention is top quartile. If you look at net dollar retention, you should really be 100 to 140% somewhere in that range. The bigger your ACVs, the larger your net dollar retention should be. The smaller your ACV is, the smaller the net dollar retention will be. And that's just the nature of the kinds of customers that you're selling to. And the fourth observation was time and usage based trials meaningfully improve conversion rates. And so if you're not using a trial that's bound by time and if you're not using a trial that's bound by usage, but instead you, you're using a seat based trial or a feature based trial, you're likely under optimizing your conversion funnel. And so it's worth testing one of those two shorten your trial length. If you're using a time based trial, there's no meaningful difference in conversion rate from a 7 day trial to a 14 day trial to a 21 day trial to a 30 day trial or longer. And so like we talked about before, capitalize on that point of maximum intent and get the salespeople to call as quickly as possible. Do hire salespeople to call leads. Again, you're 4xing the amount of you're nearly quadrupling the conversion rate and the effectiveness of the funnel. And so at certain price points where it makes sense, it's important to hire salespeople to call them. You really should be aiming that number should be a 4. Aim for 4% or better unassisted conversion. Aim for 15% or better assisted conversion on your sales question. Activity scoring in enterprise what we saw in that chart was in the majority of cases, particularly at price points above 5k, activity scoring is actually negatively correlated to conversion rate. And so it's a bad indicator. And then the last is test requiring payment. You got a two and a half increase in your conversion rate when you have payment. And you got to figure out what is the trade that I'm making when I'm actually I'm getting much higher quality of lead, but I'm also reducing the size of the top of my funnel. So those were the conclusions. I'm thrilled to share this with you all. I'd like to thank again Jason and this ASTRA team. I'd also like to thank my colleague Patrick Chase, who put an incredible amount of time into this survey and to all the companies who participated. We'll be distributing the raw data. Everybody else will be sharing the aggregate slides and future insights on the blog. But I hope you all have a wonderful conference. Thank you for sharing the first session of Tuesday with me. And I hope these data help you figure out how to optimize your funnel, maximize your conversion rates and build an incredible SaaS business. Thank you very much.