Reassessing your metrics with Earl Grey Capital's Matt Sornson
Reassessing your metrics with Earl Grey Capital's Matt Sornson
Matt Sornson:
I think the thing that founders are typically uniquely good at is they have all of the context. They understand so much more of the context of the product, the context of what other earlier customers have done, what all the evolutions between what a customer looked like on month three versus year three. And the thing that we identified when we put founder attention in was really that usage didn't equal value.
Peep Laja:
I'm Peep Laja. I don't do fluff, I don't do filler, I don't do emojis. What I do is study winners and B2B SaaS because I want to know: how much is strategy, how much is luck, and how do they win? This week, Matt Sornson, cofounder of Clearbit. He spent seven years across multiple leadership roles at the company and now serves on the board. His new day job is being a partner at Earl Grey Capital, an early stage venture fund. In this episode, Matt breaks down five key lessons he's learned throughout his career. We discuss customers hacking your product, how to scrutinize your metrics, and the benefits of implementing an outbound sales strategy. Let's get into it. Your first lesson learned is if you build it, they actually will come.
Matt Sornson:
I think that in developer tools, especially like API first developer tools, if you build it, people at least tinker with it. And early Clearbit, the very first year, one thing that we did was we would launch a new product, I'll put product in quotes here, every single month. So we had our core API, which was started as the person API, so data about people. You put in an email address, we'll give you back their role and seniority and what company they work for. Then we launched the company API, which was really just the end of that payload, which was where that company was located, what their legal name was, how many employees they had, et cetera. Then we launched the logo API, which was just one attribute of that company API payload as its own small API. And we did auto complete API, watch list API, risk API, Clearbit Connect, which was a Gmail extension, Clearbit for Salesforce, which was a Salesforce extension, a free edition of that Salesforce extension and on and on.
And I think in that first year we launched 11 or 12 products. So we launched them on Hacker News, on Product Hunt, on any sort of list aggregator where developers were hanging out and we got, over that year, 60 ish thousand people to come and create a Clearbit account. Did we get a ton of customers out of those exact marketing activities? I have no idea. We weren't set up to track it very well then. The way we were measuring the impact on revenue was incredibly basic, but it helped us build up awareness and a pretty large audience really, really quickly.
Peep Laja:
So launching something almost every single month became your marketing.
Matt Sornson:
It was pretty much the only marketing activity we did. We launched things and we told people about it and that was it.
Peep Laja:
Do you think this strategy would also work today?
Matt Sornson:
Well, I think the strategy works today. I think if you're a more sophisticated later stage company, it's probably not enough. Unless you're able to consistently increase the places where you're communicating that launch, it's probably not enough. But I would say if you're an early stage company, especially if you have an API product or an API first, then this, I think, is a great technique.
Peep Laja:
A study from 2011 shows that the most successful established businesses earned 36% of their revenue by launching new products and offerings. Think Apple, always launching, car manufacturers are always launching. A launch gives you a legit reason to reach out to people, to chat about yourself, and people like it. It plays on curiosity. What's new? Product Hunt is a clear example of that. New is a reason for people to pay attention. In today's ultra-competitive world, you can't maintain high reach and repetition without insane costs, but you can do lightning strikes. Get people focused on your launch even for a moment. If you're not regularly launching stuff, you're missing out. Your next lesson learned building Clearbit was trust the hackers.
Matt Sornson:
This was something that we experienced a lot of early on at Clearbit; people basically hacking Clearbit. And what I mean by that is finding ways to get way more out of the product than they're paying for: using the API, finding ways that they could use partial API and not hit quota limits, all sorts of things. My favorite example of this is this guy Yom Cobain, and we found out that he'd been using our person API, but not returning email addresses. So basically there was a way that you could use the API that we didn't count as an API request and he was doing it to tune of 500, 600,000 extra calls every single month. At first we're like, "Holy shit, we have to go shut this down and charge this guy more. But instead we just reached out to him, met up with him, learned what he was doing. And what he was doing was a very, very cool way of combining these data sets.
So I think my lesson learned there was don't shut down the hacker. When you find that person that's abusing your product, find out how and why. I think some percentage at the time that's going to be a huge product unlock for you, and it was for us. And with Yom, I think that happened two or three times.
Peep Laja:
People using your product in unexpected or unconventional ways, I think, has been the story of a lot of successful companies. eBay comes to mind as one of them, and I've certainly seen this also with Winter. I feel like first reaction is always like, "No, no, no, that's not how you're supposed to use that." Then it's like, "Well, why are you doing that? Oh yeah, actually our product doesn't enable you to do this straight so you need hack your way around."
Matt Sornson:
Or you're doing something completely different. Your job to be done is not the job to be done we built the product for, but that might be an opportunity for a new product service area. I will say there are many with Clearbit and many data products or even just infrastructure API products, there are going to be the abusers who are just trying to get more out of the thing. And that that's an opportunity to charge them more money.
Peep Laja:
If you have customers who have gone out of their way to optimize your product to better fit their needs, then you're getting some serious insight into what you've missed. It also confirms, with the right tweaks, people are getting real value out of what you're offering them. Here's ScreenCloud's Mark McDermott talking about how hackers helped him go upmarket with his product offering.
Mark McDermott:
We had these larger customers who were in there hacking around in the product and liking what they saw and they were kind of pulling us up and opening up our eyes, really, to the greater potential that we could serve. And so it was really more of a moment of enlightenment; and it wasn't just one customer, it was several. And once you start to see a few happen, there's just a point where there's all the data in the world can tell you something, but it's really just your gut is starting to say, "Look, I'm having these same conversations with these big customers, but there's work to be done."
Peep Laja:
Your listen learned number three is an interesting one: customer success cannot be delegated.
Matt Sornson:
I could absolutely be wrong here, but the thing that we learned in about four different customer success leaders at Clearbit over probably five years or maybe a little bit less than that was we couldn't delegate customer success successfully at Clearbit. And what changed and when we were able to really get our hands around customer success, at least a bit better, was putting founder energy and effort into it. And now I don't think that's going to be everyone's situation, but I think taking your eye off of customer happiness, customer success, even support, is a really bad idea early on. And I think it's compounded if you are a developer tool, API first infrastructure type product where the usage of your product doesn't directly correlate with the success of the customer.
So I'll give you an example here with Clearbit. It was a data API, you're sending API requests through to enrich your Salesforce, your CRM, your market automation, whatever it might be. And we would've customers that would have 80, 90% usage every single month look super, super green 'cause they're hitting the API. And then you go talk to them and something is messed up in their integration or the way that they're mapping those fields to somewhere else, and they haven't seen the data in five or six months. They're super unhappy. And I think an API product just can really obfuscate the value customers are actually getting because usage does not equal value. And that's something that was very hard for us to identify a Clearbit and has taken the combination of a lot of exec attention, founder attention, and also product attention to get us closer and closer to the place where we understand how successful a customer is based on their usage of the product.
Peep Laja:
So when you say it cannot be delegated, what stage of the company are we talking about? Is it first million in revenue, first 10 million?
Matt Sornson:
We tried to delegate it probably around 1 million is when we brought our first customer success person in. We tried to delegate it four times between 1 million in ARR and 25 million in ARR, and continued to not be as successful as we wanted to be. I think one other piece there is we had picked a market where we really liked the innovators in that market. The top 1, 2, 3% of our customers were super interesting, doing crazy fun stuff with the product. The rest of the market was doing very straightforward data enrichment, like plug it into Salesforce, enrich my account contact records.
And we were a little bit lazy and didn't have as much fun on conversations with those customers. So we didn't have that, which I think is a second learning here is customer success is hard and shouldn't be delegated early; and make sure you pick a market where you love your customers; and if you haven't, you better learn how to love those customers and love spending time with those customers.
Peep Laja:
So those people, customer success people that did not work out, what did they do wrong or what did they not do that only a founder can do or could do?
Matt Sornson:
Yeah, I'm sure it's not only a founder can do, but I think the thing that founders are typically uniquely good at is they have all of the context. They understand so much more of the context of the product, the context of what other, earlier customers have done, what all the evolutions between what a customer looked like on month three versus year three. And the thing that we identified when we put founder attention in was really that usage didn't equal value. It took us a long time to get there, but when we really dug in and started spending time with more and more of these customers, we found out that usage didn't equal value. That the top few percent of customers that we had been spending a lot of time with, the rest of the customers didn't look like them, didn't think about value the same way, weren't looking at the same success metrics, didn't know how to define success metrics.
And I think we just got way dirtier way quicker. And looking back on that, haven't spoken about this before, so this might not be the cleanest, but one thing that's hard is when you hire a CS leader, you give them metric goals that they're goaled against, and asking them to tell you that those goals are wrong and they should be doing something else is hard. We would set the goal of you should think about success through usage. We want to get every customer to 70% of their plan usage. That's how we're going to know that they're successful, but that actually wasn't the metric that mattered.
Peep Laja:
So it's being a person who challenges that, "Hey, this is not what we should be measuring or it's only telling a part of the story."
Matt Sornson:
And someone that's willing to ask those five why questions and keep pushing and keep pushing and keep pushing. I'm sure there's customer success leaders out there that can do that, but we didn't do that early on. We took our eye off the majority of our customer success and just spent our time on what we thought of as the top few percent.
Peep Laja:
It's so easy to stop questioning what should be done to begin with and just go through motions, reporting the metrics you've always been reporting on. When the measure becomes a target, it's ceases to be a good measure. Too often companies get stuck optimizing for a metric and stop questioning whether that even made sense. They start prioritizing 5% improvements over 100% by tinkering with small details. Pause and question. But what does, when a measure becomes a target, it ceases to be a good measure, even me. Here's author Brian Burns, host of the Brutal Truth About Sales Podcast to explain.
Brian Burns:
Leaders, you've got to learn Goodhart's Law. What Goodhart's Law says is, once a measure becomes a target, it fails to be a good measure. And we see this in life all the time: number of calls and sales, number of demos, number of proposals. What ends up happening is you think they're all equal and the person doing it does what's ever easiest for them and it becomes gained, so it no longer has any significant impact. You get just the opposite result of what you want because people are spending time gaining that instead of doing the real end result that you want, which is selling and getting new clients. So you got to learn this law and get away from any single metric.
Peep Laja:
Your lesson number four is about pricing, that it should be based on complexity, not usage.
Matt Sornson:
I'll paint a picture here. We're like year one at Clearbit where our pricing page is very simple. It has four columns in it, like many pricing pages do: basic, beginner, medium, advanced packages; enterprise package, contact for enterprise pricing. Each one lists our, at the time, four APIs: person, company, watch list, risk, and then how many API calls you get. So for person you get 50,000, for company you get 50,000, et cetera. And that pricing was between the end of year one and the end of year four, we close to 50x our per call pricing and completely split out all of these different APIs that had been packaged together. I'll tell you why. Because pure usage does not equal value. And what we found is the complexity of usage indicated the value a customer is getting way more than the volume of usage. And again, this is semi specific for API products, not completely. But really, there are some very small companies that are going to use a ton of your product and not necessarily get that much value or have the ability to pay for that usage, or pay for more.
There are very large companies that are going to use a tiny, tiny amount of usage of API calls, but get a ton of value because they're using it in a really complex and important way. Really interesting example of this, Dropbox was an early Clearbit customer. They used us for one attribute, a company type attribute. And really all they were trying to determine is, is this current signup a personal website or some sort of company entity? And for them, they were only using that one data attribute. They weren't touching any of the other data, their usage wasn't that high. That was incredibly valued because it drove a bunch of complex business decisioning for them, and for many of our other customers, that was useless data. So charging them the same that we were going to charge someone else for that attribute made no sense.
Peep Laja:
So really it's about value-based pricing, how much value the customer is getting?
Matt Sornson:
And it's about value-based pricing, but I think the better corollary to value based pricing is complexity versus usage. Especially if you have a self-serve motion, you're not going to be able to price each customer individually, and you need some sort of way of guessing at how much value someone is getting. Complexity is the best way to do that. So how many different places are they using the data? Are they combining different parts of the product? It's challenging to do, but for API products there's a couple ways of doing it that are kind of shortcuts.
Peep Laja:
So in Dropbox's case, it sounds like the actual application, the way they were using it was simple. Just to value, how would you figure out that you should charge them more in this case?
Matt Sornson:
Yeah, so it really wasn't all that simple because they were using it directly via API. They were using it in Salesforce, they were using it in Marketo, they were using it in their analytics products because they used that same funnel, that same personal or business in so many different places. It was so important to the business that they used it everywhere. And so being able to charge them based on, okay, you're going to use it on your website, you're going to use that in CRM, you use that market foundation and thinking of those as different, as more complex ways of using the data.
Peep Laja:
Would you then figure this out in a sales conversations when figuring out how much somebody should pay? I imagine that's easier way rather than on your website it's like, "Oh, if you have this usage pattern, that's the price."
Matt Sornson:
Totally. I think yes, occasionally, there are ways you can charge based on number of integrations. I think Snowflake does this particularly well, where they have a bunch of different levers on how many places you're going to send the data. They charge based on how much you're writing out, how much you're putting in, and how many different places you're sending.
Peep Laja:
Pricing based on complexity and not usage is the next logical step in value-based pricing models, and value-based pricing is going to be the norm. Pricing by seat might be a value metric sometimes, but often more seats does not add more value. Job boards charged by post, but that's not the value is. Might get zero applicants, zero value. Switching to value-based pricing is not easy. Sometimes value is hard to measure or it's hard to decide what the right metric is. There's also a thing that people are used to be charged a certain way, but what even is value-based pricing? Here's Patrick Campbell of ProfitWell and Paddle to take us through it.
Patrick Campbell:
So value-based pricing essentially is pricing based on the value that your customer sees, and aligning with that value as close as humanly possible. And the reason it's superior to costs in your competitors is that your customers don't care about your costs, they care about their costs. In addition to that, your competitors, you're assuming they've done their homework, which all data indicates they haven't. And in addition to that, you're also assuming you're selling the same product to the same customer, which is probably not the case. And so these data points are important, don't get me wrong, but they're tertiary inputs into basically putting your pricing together. And if that's not convincing enough, what we found is that those companies that are deploying a value metric are typically growing at 2x the rate as those companies who are only using feature differentiation. And so if you take away one thing from this particular section, go home and figure out what your value metrics should look like because it bakes expansion revenue right into your revenue model, which is great.
Peep Laja:
Your final lesson learned is about outbound sales, to start outbound earlier than you should.
Matt Sornson:
What we felt at Clearbit and what I see still at a lot of the companies that I'm investing in today is everyone wants to be bottoms up PLG, and outbound is not a dirty word, but it's not something that most founders want to talk about early on or want to do. I'll give you two examples; I'll give you the clear example and then another company that I'm pretty close with. In Clearbit, we started outbound on year four, and a year and a half later, outbound was still less than 5% of pipeline. Today, outbound, I think, is about 20%. We would love it to be 40 50% 'cause outbound is a very scalable, predictable motion, especially as the company gets bigger.
I'll give a counter example of a company in the HR space that's pretty well known. They started outbound from day one. They decided they were going to be outbound, in some ways, before they even built a product. They built their MVP and then they started just ripping outbound with six SDRs, two marketing ops data people, and the founder CEO writing the outbound email copy. And they did that for a year and a half; it was their only marketing activity; they didn't have a marketing team; they were just sending outbound email and learning what message resonated; learning what product features and what product ideas resonated.
And because getting someone to respond to an outbound email is probably the hardest first hurdle that you can think of, you really, really are forced to focus on the pains that get people to respond. And for this company, they were able to get outbound working where they were able to drive real consistent pipeline. And not only know what message was resonating, but also what product features people were most excited about in an outbound context. They used that to build their website, they used that to build all their marketing materials as they brought in a marketing team, they used that to plan a product roadmap, and they were able to go from 10 to 100 million, I think, in about 18 months on year three because they really, really nailed this in the first 18 months.
Peep Laja:
So the prerequisite here is that there needs to be a severe pain that the prospect is feeling?
Matt Sornson:
If we go back to the basics, it's like you really need to identify that pain. I think outbound is a particularly effective way of identifying pain early. I think with bottoms up or freemium, you can end up with something that's a vitamin, that people are taking or trying or using. But if outbound is the way you get people to respond, that's more likely to be a real pain.
Peep Laja:
And what did outbound program look like at Clearbit?
Matt Sornson:
Emails, calls, LinkedIn, Twitter, the whole gamut. I would say it's a relatively standard outbound playbook for the modern SaaS company. I think the unique things that Clearbit has is access to some cool intent data assets that were very useful. Our most effective outbound loop today is identifying companies that are visiting the website that are not already accounts in Salesforce, or their cold accounts not assigned to someone, and then creating those as new pipeline.
Peep Laja:
To your portfolio companies, when you advise them, do you tell them to start outbound on day one?
Matt Sornson:
As always, it depends on what the team strengths are. If the team is not going to be successful at outbound, they shouldn't start there. Start with things that you're good at, or if you have specific leverage somewhere else. But I do recommend people think about outbound way earlier than they typically do. Think about it year one, year two; maybe start your test, your trials. For more enterprise-y companies, I really think outbound from day one is really smart.
Peep Laja:
A lot of sales leaders say that if your ACV is below 15K, the numbers don't work. Is your experience similar?
Matt Sornson:
No, I think Ramp is a pretty good example here. Ramps ACV was, I think, 5 or 6K in the beginning for their SMB customers. They went very hard into outbound for SMBs their first two years and then expanded with those companies over time. If you have a land and expand motion and you think that 15 can turn into 50 or 100, then absolutely it's worth it. If it's below 15 and it will stay below 15, then you're going to need automated outbound versus like SDR-led outbound. But I still think the plays can work.
Peep Laja:
So what are Matt's top winning lessons? One, never underestimate the importance of customer success.
Matt Sornson:
I think taking your eye off of customer happiness, customer success, even support is a really bad idea early on.
Peep Laja:
Two, pricing should be based on complexity, not usage.
Matt Sornson:
Pure usage does not equal value. And what we found is the complexity of usage indicated the value a customers getting way more than the volume of usage.
Peep Laja:
Three, starting outbound sales earlier on is an advantage.
Matt Sornson:
Outbound is not a dirty word, but it's not something that most founders want to talk about early on.
Peep Laja:
One last takeaway from Matt.
Matt Sornson:
Make sure you pick a market where you love your customers. And if you haven't, you better learn how to love those customers and love spending time with those customers.
Peep Laja:
That's how you win. I'm Peep Laja. For more tips on how to win, follow me on LinkedIn or Twitter. Thanks for listening.
