Product Led Growth: Learnings from a self-serve move
Having been inundated with posts and podcasts evangelizing product led growth, free trials and self-service as the newly discovered holy grail of SaaS growth, I spent some time researching the kinds of products and pricing models that are best suited to this acquisition model before making the jump on our own product. This post contains my learnings.
The typical single digit pricing, per seat based models that worked tremendously well for giants such as Slack, Trello and Twilio have been discussed and analysed a fair number of times, but that is not the model that we are operating under. I found it much harder to find tales from the frontlines from people that had adopted a self-serve approach where the pricing model was not per seat and where the ARPU was typically in the three or even four digits per month. Surely if the general trend in SaaS purchasing was moving towards try before you buy, the advantages for the buyer should translate regardless of the pricing model.
In this post I will go over some of the points I discussed with various people in SaaS, some of my own practical learnings along the way and questions that remain unanswered that you may want to consider before embarking on a similar journey with your product.
While my personal experience has been one of adding a self-serve model to an existing product, the majority of the following thoughts and considerations would be equally useful for someone designing a product from scratch.
Consider your audience
The current accepted narrative on purchasing SaaS seems to be: as we become more experienced evaluators and buyers of SaaS, what was historically mostly a drive from the developer community has now begun to translate over to other professions that are as comfortable using a new software product as any developer and don’t need someone holding their hand in a Zoom demo to help them during their product discovery phase.
While this is definitely happening, it is also true that this behaviour has not reached critical mass yet and there are still numerous factors to consider before expecting your product to sell itself. The market you are in, the type of product, and who your buyer is all still play an important role in how receptive your audience will be to an approach that requires them to discover your product on their own. The benefit of having someone answer all your questions for you in a 30 minute demo instead of wading through a potentially complex product for the next few hours is still very real. How quickly can you get to “Wow” ?
Some products lend themselves more easily to demonstrate their core value to the user without requiring very much from them at all, making it easier to guide them to “wow” where they can hardly contain their excitement as they rip their credit card out to subscribe to your product.
Others are not so lucky and by their nature require significantly more effort from the user to be able to see any value at all, making it that much harder for a product designer to convince a user of the product’s benefits in the short time the average free trial user is willing to give up.
“What would it take to get a new customer to a non-empty first screen?”
When speaking with Shaun Clowes (SVP Product @ Mulesoft) about this, this is how he framed the question. If your product, like ours, relies on data from other sources to provide value, how easy would it be for you to get that data? If your answer is “easy, we’ll just get them to authenticate to the external service and populate their account in a single click”, think again about how much trust you would have built up with the user by this point, and if that’s sufficient for them to be comfortable with handing you the keys to their company’s valuable data; probably not. And that is assuming that the trial user actually even has access to the data that you need: in our case we need Instagram data, but it was unlikely that someone from the eCommerce team, our typical buyer, would have access to the company’s Instagram account.
Pre-populating your product with demo data for the user can help in overcoming some of these hurdles, but you need to consider if demo data is going to pack enough of a punch to push the user across the hypothetical free trial finish line.
Pricing, Trust and Risk
I started this research with pricing being the most intriguing part of the puzzle for me, and it turns out there was no clear general conclusion that I could draw. In some industries there might actually be a clear monetary limit above which self-service starts becoming less likely due to restrictions such as credit card limits and procurement policies (I’ve heard $5k ACV being the cut off point for some SaaS products), and others where this rule just does not seem to apply at all (Atlassian has self-serve customers paying $100,000+ annually).
Trust definitely plays a significant role here as well. The bigger SaaS companies I mentioned previously that experienced massive growth with their use of free trials all had per seat prices in single or low double digits. This is almost like an extended trial, where you start by buying the license for a few seats to see how well the product solves your need at a risk of close to zero (low per-seat cost on a monthly billing plan), building trust in the product and the company before rolling it out to the rest of your organization as quickly or as slowly as you desire. I’m not sure if it would be possible for companies like Atlassian or Slack or Salesforce to have self-serve clients paying six figure license fees if they did not employ per seat pricing enabling this kind of trust building.
Non-SaaS products and services that cost tens of thousands, such as a first class flight ticket or maybe even a Tesla, are being bought online everyday without the need for someone to speak to a human at the point of purchase. These companies have all managed to build up enough trust with the consumer for this to happen and maybe the question is not where the highest ACV is for you to be able to self-serve your SaaS, but rather how you can build up as much trust as possible in the least amount of time with your potential buyer ?
Free trials as lead generation machines
At this point I’d like to re-enforce the distinction between free trials and self-serve models: one can exist without the other. A free trial by definition does not require the user to be able to convert to a paid plan on their own, as this can still be done by a sales team and might be a more suitable approach for complex enterprise products where the trial serves solely as a lead generation tool.
A good example here is Salsify (who curiously enough have recently removed their free trial option), a SaaS focused on Enterprise eCommerce brands, that do not offer users the option to upgrade from within their free trial on their own. Due to the complexity of their product the initial results from introducing a free trial were actually counter productive as users could not get to see value quick enough (HubSpot had a similar experience), and therefore would not easily convert into paying customers even with a sales assisted model. Adding demo data and app cues significantly helped ensure that users understood the product and had an enjoyable enough experience to want to learn more. In this case, the free trial was used purely as a lead generation machine that fed the sales team with highly qualified leads that were already impressed by the product. The added benefit here is providing your champion with a means to demo and evangelize your product internally post-demo.
Product complexity is another critical point to the success of a product led growth strategy. As mentioned before less complex products lend themselves better to a self-service model, due to the ease and speed at which users can get the value they expect from the product. The more complex the product, the higher the chance of a potential customer requiring a human to help them evaluate and make the purchase.
When speaking with Brian Balfour (ex-Head Growth @ Hubspot), the biggest problem with HubSpot’s free trial in the early days was the complexity of the product and therefore not being able to get to “wow” quick enough. Slack, purely by being a product of significantly lower complexity, would probably not be facing the same issue.
While you can’t suddenly turn your complex product into a simple one, and in many cases you probably would not want to, besides adding app cues and demo data to ease the onboarding process, the idea of an “intro product” is something that has worked well for some companies ..
The Intro product
If your main product is too complex to easily do a free trial with, consider a reduced version of your product. A small subset of your main product, attractive to your target market, no setup complexity and lends itself to being easily made into a freemium product. Upselling to your main product in-app, or through a sales team, could then be an effective source of qualified leads.
I see a difference here with tools that are often clubbed together in the “engineering-as-marketing” category (such as the used to death example of HubSpot’s Website Grader) where the tool is typically a new product altogether requiring significant additional marketing resources to acquire users for it. The intro product is not actually different from your main product, just a simpler version/subset/first step of it, and therefore does not require additional marketing efforts as your messaging and value proposition stay mostly the same.
An example of a good intro product is Ahrefs backlink checker: A valuable, although limited, “standalone” tool offering an insight into the power of Ahrefs and enticing the user to buy the primary product now that they have been convinced by the quality of the data. It serves as a first step into the complete Ahrefs product suite.
As is common with development teams, and largely due to the unforeseen complexity of many changes that seem quick on the surface, the amount of resources required to introduce a free trial/self-service component is often underestimated.
Integrations with payment gateways, a solid guided tour solution to help your users see the value they signed up for as soon as possible, the possible additional load on your resources (your technical infrastructure, but also your customer support team) as well as the continuous maintenance as your product matures, are all time consuming and costly. Companies with smaller development teams would be wise to consider the effect this model would have on their teams and pace of development before jumping into it.
A unique example here is the market leading SEO tool, Ahrefs, that charges $7 for their 7 day free trial. While adding friction to the top of their funnel, this no doubt significantly reduced the load on their resources from users that were unlikely to become paying customers.
Fake it until you make it (or making data driven decisions)
As I’ve laid out in the rest of this post, there are no hard and fast rules on what kinds of markets, industries and professions a product led growth approach would be successful in. Given that, doing a significant amount of research and data gathering on how likely your customers would be to buy a product like yours on their own before you embark on this journey is something I would highly recommend.
One approach is to test the receptiveness of your leads with a far-from-automated solution, such as putting up a free trial form, where the user will have to wait for their demo account to be manually setup. You will not win any awards for user experience, however it will very quickly demonstrate how keen your leads are on buying from you in this way before you delve into what will most likely not only be a large ongoing development project, but a mindset shift within your organization.
I hope this has been helpful. Feedback appreciated!
Thanks to the following people that helped me with my research:
Gokul Rajaram (Caviar Lead DoorDash)
Jessica Qu (ex product manager At Salsify)
Shaun Clowes (SVP product MuleSoft, ex Head of Growth Atlassian)
Russ Heddleston (CEO Docsend)
Justyn Howard (CEO Sprout Social)
and via Reforge:
Stu Stein (Product manager Peerspace)
Brian Balfour (CEO Reforge, ex Head of Growth HubSpot)
Claire McGregor (Co-founder Appbot)
Interesting content on the topic: