The world’s been turned upside down in 2020 and it’s left many businesses wondering when they should consider scaling again - this question’s one I’ve been pondering for years and something my colleagues and I are trying to introduce a little more scientific rigor too - when to scale, where, and how?
In this article, I’ll explain what we’ve come up with when it comes to adding quantifiable measures to identifying product-market and go-to-market fit - the combination of the two we suggest signals a business is ready to scale, and I’ll go on to explain where and how.
My name's Mark Roberge and in this article I'm going to talk about the science of re-establishing growth.
We've gone through an unprecedented journey in the last few months of trying to reorient our roles and our businesses. I know it's been a struggle for everyone and hopefully, we're feeling a little bit of stability now. I know a lot of people are thinking about should we grow again? Or when? And how do we think about that?
When do you scale?
That's something that I've been thinking about for the last couple of years is this question of when do you scale? How fast? Where? Such important questions for a startup, for any company, and waiting too long has a pretty negative effect of leaving an unnecessary window open to competition, and perhaps even missing the window for the opportunity.
Going too fast, I think kills many startups. I think this is a really important question and I just want to challenge us a little bit on, can we bring a little more data and rigor to this strategic decision around when to scale, where, and how?
#1 Product-market fit
When I ask people when should you scale? Most of the time people say when you have product-market fit. I agree with that.
What is product-market-fit?
I think the challenge is when I ask 10 people what product-market fit is I get 10 different answers. Kudos to Eric Reese for possibly creating this term, or Steve Blank, Marc Andreessen, and Shaun Ellis for adding substance to it.
But it still amazes me that for such an important term, an important decision, we're still pretty subjective in how we define it. I was just on a webinar a couple of hours ago, and I asked a very smart moderator what she thought it was, and she had a good answer - ‘demonstrating a new innovation that works within a big market’.
It's actually pretty good because I would say most people say it's because of revenue, and I don't think that's good.
‘Product-market fit is about sales’. No, not in my opinion. I think that's message-market fit. You know how to get someone on the phone, understand their needs, and deliver a pitch that says "Yeah I'll buy". But product-market fit has to be more around delivering the value.
All you've proven with sales is you know how to sell - I think that's a little dangerous definition. When we get into a really new innovation proven in a big market, I think that sounds better it's just a little squishy, a little subjective, a little qualitative.
We've seen some people say survey your customers - and if 50% say they love it you’ve got product-market fit, but surveys are notorious for false positives and that can really mislead us down a rabbit hole of scale at the wrong time.
If people fill out a survey, they want to say nice things, I don't know if it really represents their true feelings and behavior. When I think about product-market fit, if I had to choose one metric to base it on, I would say it's customer retention.
Customer retention is the best quantifiable measure of product-market-fit
I think a lot of folks understand how we're measuring customer retention these days, that world-class is 90% annual customer retention, 100% annual revenue retention - that's world-class, I think if you have that you have product-market fit for sure.
The problem is whether we're a small company growing up, or a big company that's trying to re-establish if we have it, we don't have enough time to know if this truly surfaces.
If we acquire a bunch of customers this quarter, the long term retention rate of those customers may not serve us for over a year, we don't have that much time.
What I don't see enough of us do as entrepreneurs is seek out a leading indicator of customer retention.
Define a leading indicator to customer retention
When we can consistently achieve the leading indicator of customer retention, that's the best I could come up with as a quantifiable measure that surfaces relatively quickly on what true product-market fit is.
With this, I'm trying to codify good definitions of a leading indicator of customer retention.
If P% of customers achieve E event(s) within T days.
What are these variables?
Let's look at some examples.
Industry examples of customer retention leading indicators
Slack
If you read up on Slack, sometimes Silicon Valley calls it the 'aha' moment, 70% of customers send 2000 team messages in the first 30 days. Great. That happens. That's the best definition of product-market fit I could come up with for Slack.
Most people that use it within 30 days are using it in a way that is aligned with their value prop. I imagine of those that actually send 2000 messages in the first 30 days, a lot of them will stick around forever. I would say if they don't probably a lot of them will fail, will churn.
It's a good definition.
Dropbox
I think if you look at the 'aha' moment research, 85% of customers upload one file in one folder on one device within an hour. Good.
Hubspot
HubSpot I know what it was because I was there while we're doing it - 80% of customers use five or more features out of the 25 features on our platform in the first 60 days.
Hopefully, that gives you a sense of what we should be seeking.
In terms of these individual variables:
- P% - I see it usually somewhere between 60 and 80%. Don't get caught up in the exact number, just the majority percentage. Pick a number to set a goal for the company, but the majority.
- T days - It really depends if you're selling pencils or jets. Dropbox within an hour, HubSpot takes 60 days, something like selling to enterprise, it could take six months for the lead indicator to surface. So big enterprise deals - many months. Super transactional PLG situations - hours, maybe a day.
- E event(s) - There are a handful of things here.
- One, it's not subjective, it's binary, yes or no. Not like 'the CSM said ‘they're in good shape', that wouldn't be good. Something that we could observe, it's usually around product usage or product setup, or maybe the performance of certain metrics that indicate their ROI.
- Two, it can surface in a short period of time.
- Three, it can be instrumental because I'll show you later how we're going to want to measure this forever to know if we're scaling too fast.
- Four, ideally correlated with our unique value prop. HubSpot is a perfect example of that, we competed against point solutions. If you just wanted to blog get WordPress, if you wanted to do social media get Hootsuite, if you want to measure the effectiveness of your funnel, get Google Analytics. But if you want the powers of having all that in one, then HubSpot is your best solution. The fact that the E event - using five features out of 25 features matches one - was super correlated with their unique differentiability that's powerful because we're going to align the whole organization around that.
The way we look at it, to know even sooner, do we have product-market fit is to organize our customers by customer cohorts when they were signed up.
Instrument customer acquisition cohorts
This particular company if we make up the fact that this is Slacks numbers, they signed up 24 customers in the early years in January, and 3% had hit the early indicator of retention after one month, not good, and 33% after three months, not good.
You can't tell me that you defined this event that represents the value your product delivers and only less than 50% of the customers achieve that in the first nine months. You don't have product-market fit, but they did a bunch of changes in this example to the product, the onboarding, how they sell, the type of people they're going after, etc. and look, by September, they signed up 50 customers.
Sure, a small percentage of them, only 12%, in the first month achieved the lead indicator of customer retention. But by month two 68% achieved it, by month three 75%.
Somewhere around there, I think they hit product-market fit. I would say this is the quickest statistical representation to identify when that happens, which is so key for our ability to learn fast.
Over time, verify the leading indicator correlates with customer retention
This is just an example of how over time, you can prove whether it was right.
Strong correlation
Literally a year later, hopefully, you've gone into the scale mode now, you didn't wait for this to come about, but you can verify, that this particular company signed up 68 customers in Q1 of last year.
Somewhere between 12 and 18 months ago, they signed up 68 customers, all of them are over a year old. They're between a year and a year and a half old. And 55 of them had achieved that early indicator of customer retention in the first 30 days, and 13 did not.
As we evaluate those cohorts a year later, turns out that those that did hit the early indicator of retention, 93% of them are still here. For the ones that didn't only 39%. Bam, you crushed it.
You know something very important about your business, which is what can you see in the first week, month, two months, that if that happens, the customer will probably with you forever. If it doesn't, they will churn.
Weak correlation
The bottom example shows you didn't have it. The retention rates on those two buckets are very similar so try again.
You didn't waste the last year, if your lead indicator was to get your customers to set up the product very quickly, and use it, believe me, you didn't waste that year. That's not gonna hurt you. It's just you haven't quite figured out what statistically correlates, you have to dig a little bit deeper. You have that insight.
There you go, you have product-market fit if the customer retention lead indicator correlates with long term retention. It is true. And it is true if P% of customers achieve E event(s) within T days.
I know it sounds complicated, but hopefully, the way I've laid it out you can see the data, the rigor and the precision behind something like that, and how it can affect your business.
#2 Go-to-market fit
Let's move to go-to-market fit. Notice I didn't say anything about profits or scale in customer retention, all I said is figure out of the customers you sign up in a month or a quarter or whatever most of them realize the value of your product quickly.
What is go-to-market fit?
Scalable Unit Economics
Go-to-market fit is just doing that scalably. When we think about scale, especially in SaaS, we talk about Unit Economics. The same issue is true with Unit Economics - when we sign up a cohort of customers this quarter, we really won't know whether the Unit Economics are good for a year.
Defining the leading indicators to Unit Economics
We talk about lifetime value to CAC ratios, we talk about payback periods, these are the metrics that we're looking at and we just won't know for a year. So we have to find similarly the leading indicators to Unit Economics, which is I think, easier than the lead indicators to customer retention.
It's really just algebra, and I'll walk you through and if you're not quantitative, your head will be spinning at the end but stay with me, it doesn't matter because I'll show you how you simplify this in the dashboard - it's just so we can see the definition.
Here are some acronyms that we are familiar with:
- Lifetime value,
- Customer acquisition cost,
- Average contract value - how much do your customers pay you every year,
- Your gross margin,
- A sales qualified lead,
And here's our target LTV/CAC greater than three. That's where you want to be.
Now we just have to extract it back into the observations of our business that we have today to know that a year from now the LTV CAC will be greater than three. Let's break these down.
LTV
LTV i's a function of how much your customer’s paying you a year, ACV, times the gross margin percentage divided by churn. That's one way that most people use to calculate it.
CAC
The cost to acquire a customer is the sum of the cost to generate the lead, marketing, plus the cost to sell the lead, sales.
The cost to generate the lead is your cost for your SQL divided by the conversion from SQL to customer.
- If it costs you $100 to generate SQL and 100% of those SQLs turn into customers, then your marketing cost for a customer is $100.
- If instead, the conversion rate is 10%, if only one in 10 become that, then your marketing costs $1,000.
On the sales, it's really just a function of how much you pay your salespeople every month divided by how many customers they sign up, which is a function of how many SQLs they get per month times the close rate.
At the end of the day, we extract it back to this seemingly complicated algebraic calculation, and basically, it allows us to put together a business model.
An example
Here's an example like, you pump it into the formula, we're gonna shoot for average ACV of 20,000, shoot for gross margin percentage of 70%, churn 15%, annual cost per SQL of 1000, SQL on the customer 5%, you get the point, it spits out a good business.
Instrumenting the leading indicators to Unit Economics
Now we're in a position where we can instrument our business, we're all using CRMs and analytics dashboards. This just allows us to dashboard out the key elements of that formula so that knowing that we want to have an LTV to CAC of greater than three a year from now, it tells us how many SQLs we need to generate per salesperson, what the cost per SQL is, what the conversion from SQL to customer is, what the average spend per customer is, for us to generate that number.
That's the leading indicator to Unit Economics and that's what I defined as a go-to-market fit.
Now, I think we are ready to scale. This could take a week or take a month, it could take three months. But we focus the organization sequentially on these two components and this metric becomes our speedometer.
#3 Growth and Moat
Time to scale: how fast to scale
When it comes time to scale when we have product-market fit, defining in this data rigorous way, and we have go-to-market fit defined in this data rigorous way, we can now scale.
A common pitfall
Unfortunately, 95% of people who scale whether they're a startup scaling for the first time, or they are a new product in a bigger company scaling for the first time in that product line, they add 20 reps in the next month and it fails. It's really hard to hire 20 reps into a system within one month.
Establish the pace. Watch the speedometer
Scale is more about a pace, a pace that's metered by the speedometer. Let's imagine we’re a series B startup, we're ready to do this, let's hire two reps a month for the next four months, and let's watch the speedometer.
Let's watch the lead indicators of customer retention. Let's watch the lead indicators to Unit Economics. If they go red, let's intervene, fix and get back on pace. If they stay green, let's go faster - four reps a month, eight reps a month, but in a very scientific way.
Hopefully, that gives us some insight on when to scale, when we have product-market and go-to-market fit, and how fast to scale.
Set a pace, go as fast as possible without breaking the speedometer.
I just want to say most companies evaluate the pace of scale based on the P&L. I believe the P&L is a representation of what happened in your business six to nine months ago. Versus this is a representation of what's happened in your business today.
Thinking about the pace of scale based on as recent performance metrics as possible.
Where to scale
This is very applicable to the larger businesses, this particular business in the image below is around 30 million, and at this point, the average has become pretty dangerous.
Because if we just look at retention and Unit Economics, and they look good, well, usually those vary quite a bit by the product you're selling, the market you're selling into, and the channel you're using to sell.
Those are the three dimensions in my opinion on segmentation of the go-to-market.
In the above example they have two products, the current product and a new one they started selling, they sell in two markets, mid-market and enterprise, and they sell through two channels, a direct channel and a partnership team.
You can see by the numbers:
- The direct channel and the current product's doing great. Great LTV to CAC, great retention.
- The partner channel on both is suffering. LTV to CAC is too low, churns too high in the mid-market.
Separate ‘scale’ teams from ‘experiment’ teams
This analysis allows us to separate our segments into scale versus experiment versus ignore segments, which is a CEO/founder job for the business to make sure we're not boiling the ocean but also to make sure we're unleashing new growth avenues.
When we do, to not overcommit like, "Hey, guess what, we're a $30 million business and we want to go to 60 million next year. So here's my idea, let's build a new product and let's tell the board that we're going to get 10 million in the first year from that new product. While we're building it, we'll train all of our sales people about the new product".
Are you kidding me? When you were a seed-funded business, you didn't hire 20 reps before your product was done, there's no way this new product is going to work out of the gate. Your job is to learn as fast as possible by keeping these new product or market or channel teams cross-functional and small so that you can learn as fast as possible.
Then scale and feed the core markets that you already know how to scale. That's where you're gonna add your people to.
Hopefully, it gives a little roadmap of how we're scaling, where we're scaling.
Aligning GTM with the pursuit of product-market-fit
The final piece is, now that we've got a more data-intensive approach to figure out which stage we're at, which sequential stage, it gives us the opportunity to optimize our very tactical go-to-market decisions on that stage.
The type of rep we want to use in the product-market fit stage is way different than the growth stage.
The type of customer going after is different, whether demand Gen should be worked on or how we think about pricing.
Product-market fit
When we're at the product-market fit stage we want to focus on early adopters who are not asking for dozens of references but are the types of people who define themselves either as a company or a persona as an innovator who wants to tinker with new stuff and give you feedback on where it works and where it doesn't.
We talk about having a win at all costs playbook, good founders at this stage are getting on the plane to meet with $50 a month customers to get them onboarded, do an unscalable thing.
Paul Graham from Y Combinator's advice - unscalable things early. The salesperson is a mix of a product manager and account executive, the abilities of a PM are to talk to a dozen customers and sense make the patterns and then communicate them to the engineers, but the ability of an account executive is to progress a deal forward, they're confident to discuss money and objections and get to a close.
Demand Gen, pricing, compensation doesn't matter. You should have a strong enough network between you, your employees, and your investors to get at least a few dozen companies to see if you can get product-market fit.
Pricing - free is bad, but extreme discounts are fine. Just make sure they're committed, you want a chance to show the success.
Go-to-market fit
But that changes when we get to go-to-market fit stage, now we have to do it scalably we need good Unit Economics.
This is where a codified sales playbook matters. We need reps, probably four of them, to build the process. They love listening to film all day and changing the playbook to optimize and get it ready. We need at least one scalable demand Gen channel, cold calling, content marketing, channel, trade shows, whatever, we need something that if we're going to hire two reps a month we know how to feed them.
Then pricing model and comp plan - this is where it matters. This is don't lose sight of the retention. My favorite comp plan is half the payment on the signature half the payment on the leading indicator of customer retention.
It's only gonna take 30 or 60 days to defer half of it. It's not the salesperson’s job to onboard the customer. It's their job to set good expectations to make sure that it happens.
Growth and moat
Once we move into growth moat now we've started the multi segments, we have a sales management layer to reinforce the playbook on the people that we hire. I need the coin-operated rep as they coined in the sales learning curve 20 years ago, that takes the comp plan and the playbook and is ready to win.
Hopefully, this opens our eyes up to a more scientific approach to this day and age knowing when to re-establish growth. And in any day and age, knowing when to scale and how fast.
Before I go
I've actually codified this into a working document that's about 40 pages. If you want to go check that out, I'm always open to feedback. Part of this is figuring out your bottoms up growth model, if you want to download that on the Stage 2 Capital website, we have other assets like hiring your first sales leader, or just general on our blog, if you want to subscribe and check that out.
Stage 2 Capital, by the way, is the first venture capital firm run and backed by sales, marketing, and customer success leaders. I'm extremely humbled by the group that's gotten behind us.
Many C-level executives from most of the unicorns, Facebook, Yelp, Tesla, SurveyMonkey Dropbox, Adobe, Zoom, you name it, we've got executive representation from there, and I've learned so much from this group.
I'm very excited about how we continue to influence especially the scale stage of the entrepreneurial ecosystem. I've been on the faculty at Harvard Business School for five years, you can check out my publications on my faculty page if you'd like.
Finally...
Finally, I did write a book at the end of my tenure at HubSpot called the Sales Acceleration Formula, I just want to let you know that 100% of the proceeds from that book are donated to build.org, an amazing nonprofit.
There's so much necessary talk about the growing divide of inequality in our world and in this country, and Build partners with some of the worst-performing high schools in every major city and basically puts their at-risk students into an entrepreneurial program to help them find the passion of entrepreneurship with the intent to get through high school and again in college.
99% of those participants grade through high school and 85% matriculate in college, which is way above the average rates for those schools.
If you've supported the book, thank you, you're also supporting build.org. Thank you.
