Patrick Campbell doesn’t do data analysis to a PhD level, but he has always sought leverage through data for the decisions he makes. As the Founder and CEO of ProfitWell, he’s witnessed many startup founders bumping up against a common predicament: How do you make data-driven decisions when you don’t have a tonne of data?

We had Patrick along to one of our Open Source VC meetups, where he chatted to our Partner Jax, unpacking his startup pricing and go-to-market (GTM) strategies that have earned him the reputation as one of the most influential thinkers globally in this space.

How should startups think about pricing decisions in the early days when they don’t have much data?

In the early days, you’re never going to have perfect data. Instead, you’re chasing a qualitative feedback loop to help with decision-making and the tests you’re running. But you’re also using intuition, and the only way you’ll know whether something is a fool’s errand or not is to throw stuff up against the wall, try and read the tea leaves and then do it again. Then the next time you do it, you’ll have more data.

What is your framework for customer development to validate assumptions on pricing?

I start by asking a core question to 10–15 people that will help me learn about the problem. I never ask questions like “This feature is kind of cool, right?” or “Hey, what features do you think we should have?”. I ask broad questions like “What do you think about solving this problem?”. And in these conversations, I want to be a prosecuting attorney; I don’t want to lead the witness.

If the conversation is around pricing and willingness to pay, there are lots of ways you can ask this. One that I have found that’s efficient in terms of both time and money is The Van Westendorp model. The basic idea is that we as humans think about value on a spectrum. We know that a computer is worth more than a bottle of water. But if you’re in the desert for 3 days without water, all of a sudden, that value differential changes.

I apply this model in meetings by showing them the product and asking, “At what monthly price is this way too expensive that you’re not going to return my calls or emails?”. Let them give you a few numbers and then ask, “At what point is this a great deal, like you give me your credit card right away?”. Over 10–15 people, you don’t have quantitative data, but you start to have a range, so you know if you’re a $10, $100 or $1,000 product. That’s the kind of question you’re trying to answer in the early days, not if you’re a $19 or $20 product.

The next step is to go wider and survey 250–300 people. You want them to answer the 2 questions already mentioned, plus 2 more:

• At what point is it getting expensive, but you’d still consider purchasing it?
• At what point is it too cheap that you’d question the quality of it?

If you use the standard Van Westendorp analysis on the answers to all 4 questions, you usually get a plus or minus of about 20–25%.

After going through this process for one of our products, we discovered 2 big things. The first was the willingness to pay for the product from our target customers was about a third of what we thought it was. Our projected unit economics for this product, such as the customer acquisition costs and what we thought our lifetime value would be, was out of whack.

Our second finding was that there was only a 3 to 5x difference between our smallest and largest customers. And whenever you have this scenario, it’s terrible because you’re never going to get any upgrade or expansion revenue no matter what you build.

The tl;dr is that most companies should do a bit of this research as early as possible so they don’t end up building a company that wastes a lot of time and money.

How do you think about GTM channels in the context of unit economics?

‍From 2000 to 2010, the challenge was product-market fit, and the hard part was building the product. Once you had the product, you’d go find the market. There wasn’t a tonne of products or competition, so it was easy and new marketing channels were opening up every quarter.

We haven’t had a brand new marketing channel since Snapchat in 2015. Only now is TikTok emerging. This is why we’re all reinventing marketing channels from things like emails and sales to account-based marketing.

Brian Balfour was the first one to start talking about why product-market fit isn’t enough. I reference market, product and channel fit, which is: What’s the market you’re going after, to figure out what the product is, to get market-product fit. Then ask, where are those customers?

The biggest thing to figure out is where your customers are. For those going mid-market or enterprise, you will need some sort of multi-touch strategy, e.g. sales, content and referral programs.

What do you think about freemium as part of a GTM strategy?

Freemium works well because it puts the onus of conversion on your customer rather than on you and an artificial timeline. This is important because with the rise of competition and compression of acquisition channels, you need higher quality executions to bring customers in.

It used to be, “Oh, we’ll just write better content”, which is why blog posts went from 800 words to 2,000 words. But:

Some of the best content you have is your actual product. That’s what they’re willing to sign up for. And if they get value and start to buy into your brand, then there are some triggers you can pull to convert that person to a paying customer.

There are many ways you can do freemium. Take the faux free trial as an example, where you give the user 20 opens a month. They use those opens in the first 14–21 days, then they get hit up to convert, and that number resets at the end of the month.

What I recommend doing is looking at your freemium conversion on a cohort basis, rather than just focusing on the first 30 days, to see who’s converting 3, 4 or 5 months in. That really shows the power of freemium.

When should a company experiment with freemium?

Too many people read a product-led growth article and then think they should do freemium. Until you have someone focused, from a product or conversion perspective, on the metrics between free and paid, it can increase noise considerably.

My recommendation is, if you’re a B2B SaaS in your first 6–12 months, go figure out how to sell your three and four figure Monthly Recurring Revenue (MRR) deals exclusively. That’s where there are the most constraints and where you’ll really learn from people paying, for lack of a better phrase, real money.

How do you build a broader self-service offering for your enterprise customers?

Very carefully. You have to understand your company’s DNA first.

At ProfitWell, before we researched our TAM, we thought, “Let’s go get a bunch of $50 customers.” But the problem was that our DNA was really good at selling high 3-figure and low 4-figure deals. And then we thought, why mess up everything for some DNA that we don’t have?

So the first question is, should you do this? Do you realise how much work it’s going to be to shift or open up another team? Assuming you’ve done that, you want to start to learn what deals your Account Executives can handle. If you’re well-funded or have good cash flow, you might hire someone right away for them.

The buying motion and the pricing will be different, and that’s something I’d spend some time on before jumping in. And that will also help you convince your stakeholders that the strategy has been battle-tested.

What should companies be thinking about when it comes to expansion and churn?

There are two fundamental things to do.

The first is if you’re a B2B SaaS company, you need to figure out your value metric. It’s how you charge per user, per 1,000 visits, per dollar retained — whatever it is. This is so powerful because you bake expansion, revenue, and churn into how you make money. In every single metric, this is the thing that hedges and leverages you against yourself and the mistakes you’re going to make in the process of building a company.

The other thing which is underutilised by mid to late-stage companies is add-ons. Add-ons are not quite standalone products. They are additional features that they can sell to someone who’s already happy and willing to pay more based on value exchange.

Could you talk through the concept of understanding the value metric of your company?

The first thing you want to do, and it doesn’t matter if you can measure it, is understand: What is the perfect measurement of value that your customers get?

Using HubSpot as an example, the value metric is the money they bring you through the marketing software. You then want to find the 5–10 things that are the proxies for the marketing dollars they bring you because it can be hard to decide whether it was the blog post you wrote or the automated email they sent. The proxies for HubSpot may be per user, per visit, number of visits, per page views or per contact.

Whittle the proxies down to 5, and then ask your customers of these 5 ways to price the product; what is your most preferred way of pricing and what is your least preferred way of pricing from that group? And then you’ll start to see some consistency around proxies. For HubSpot, it mostly comes back to the number of contacts as it’s an anchoring point for how many people think about marketing automation products.

Finally, you measure willingness to pay. This is to ensure that people who, for instance, have a lot of contacts are willing to pay more than people who don’t have a lot of contacts. You want to make sure that value metric is linear with the number of contacts and willingness to pay.