Hi and welcome to the third episode. I'm your host, Vincent Young, and today I'll talk with Pekka Koskinen. He shares how they developed the product-led sales motion at Leadfeeder long before terms like PLG and product-led sales existed.

This interview is part of my upcoming book about product-led sales. If you're doing sales today and are wondering how you can leverage PLG, then this book is for you. For early access, visit plgandsales.com.

Now, let's introduce Pekka.

Before starting Leadfeeder, he founded three other businesses. These include a software development company called Solenor and a payment gateway company called Payment Highway. In 2004, he founded Snooby, a web analytics business that provided much of the early learnings for Leadfeeder. These businesses have all been sold, and today he’s fully focused on Dealfront, the result of merging Leadfeeder with Echobot after an 80 million euro investment from Great Hill Partners.

At Dealfront, we work together on a daily basis, building on the product foundation that Leadfeeder brought into the joint company.

Alright, Pekka, great to have you here today and thanks. We'll be talking about something that we don’t spend a lot of time on usually, which is how Leadfeeder actually got started and how you built the growth model over the years.

Sure. So, we started Leadfeeder in 2012 with my co-founders Her and Vent. Before that, we were running a web analytics company called Snan, which was a Finnish-based business. The majority of our customers were in Finland, and we were selling with outbound sales. We were figuring out how to go international.

This was around 2010, and we started to build outbound sales teams in different countries. We learned that doing the outbound sales-driven internationalization motion wasn’t easy—it was very costly. At the same time, we were seeing really great startups coming from the US, like HubSpot and others, that had this nice free trial experience. They were generating leads at a fraction of the cost that we were, and they were able to close deals better because customers had already gotten some value before they needed to purchase something.

When we made the exit from Snooby in 2012, I wanted to build my next startup very differently. I wanted Leadfeeder to be an international, global, product-led growth first company, selling with dollars and not with euros. We thought that we needed to look and feel like a US company. This meant we wanted to become really effective and efficient in how we went to market.

How we did this was that in the first couple of months, we built a prototype of Leadfeeder and just put it out there. We didn’t even try to sell—it was a free sign-up, free tool. You just signed up, used it, and gave us feedback. We started to do Google ads on it. The idea was to start validating that there was actually a problem and that we had a solution for it.

We ran Leadfeeder as a free product for a year until we then put the most interesting features behind the paywall and started validating the next step in the startup journey: are customers willing to pay for this? Are they willing to pay month over month, which is a key thing to prove product-market fit?

Back then, we didn’t have any support or sales around it. It was just pure self-service. Luckily, we started to get purchases from all over the world. We had signups from the US, Central Europe, the Nordics, and everywhere. Many of those found the product to be good, put in their credit card, and started paying us.

We, of course, developed analytics around it—tracking how much traffic we could drive from ads, how well they converted into signups, how the signup funnel worked, and the conversion to paid. Then we started to improve this.

We saw that if we could get in touch with customers—like in the beginning, we only had support, not customer access or sales—we saw that if we could talk with customers through support, the probability of making a sale was higher.

When we had a good flow of trust coming in, we started to think: would it make sense to contact these customers and sell to them? We hired our first salespeople, who began contacting the most promising trials we had.

And that worked really efficiently compared to the old model we had experienced in the previous company, where we were doing cold calling. This is how the product-led motion got started. Then we started to add commercialization steps and hired more people to call the trials. We were experimenting with what kind of trials we were conducting and how much effort we were putting into them. We were quite light touch, so if they didn’t buy, we moved to the next one quickly and didn’t force the sale. We scaled this motion for many years. Then, around 2017, when we had several million in ARR, we started to look at other channels as well.

With inbound marketing, you can only reach those already searching for a solution, but we knew there was a big group around these innovators and early adopters who didn’t yet know they needed our solution. You can only reach them through outbound. So, we also started to invest in outbound sales and channel partners like marketing agencies. These were the two ways we could go outbound: having someone else talk about us (marketing agencies) and hiring outbound salespeople to contact bigger clients. We then evaluated the customer acquisition cost compared to the lifetime value from each channel and started to put more money behind the channels with the best ratios. That’s how we ended up with a model where we have several channels serving different purposes and targeting different audiences.

You mentioned something interesting at the beginning. You first wanted to get people to use it and only then figure out if they were willing to pay for it. I know from other conversations that in outbound, you’re still doing something a bit unusual. Your outbound motion is still focused on getting people to use it first, right, instead of the traditional sales approach where outbound sells something and then gives access. How did you find ways to do something that’s more unconventional?

Yes, that’s true. Our outbound still went through the free trial phase, so there was the usage phase. This was a must-have because on the website, we were pushing the 14-day free trial so much. If we just tried to sell somebody with outbound, they would visit our website and say to the salespeople, “I want the 14-day free trial before I decide.” It was hard to avoid that route, so we decided, let’s do that with outbound too—have them go to the trial. We knew that if somebody took a trial and installed the tracking script on their website, we could demo the product with their data—website visitors, customers—and that would be way more powerful. We had tested this and saw what the conversion rates were with clients where we demoed their own data compared to us demoing our website data or some demo site. It was totally different. We could already set up customer accounts, searches, or filters during those demo calls, and that proved to be very powerful. After the demo, the customer had access to the tool and could continue from where the discussion left off. That was really powerful. We just called them afterward and tried to close the deal because the product had already proven the value.

You’ve already touched on some of this, but what were the key moments on your journey, especially regarding the go-to-market?

I think this is related to how you build a startup. In the beginning, there was the validation phase—does this solve a problem? For that, I advise many startups to just give the product out for free. You can always upsell them later. If you put a price on it, you’re not losing revenue because they didn’t pay in the first place. The second phase was picking the most likely working go-to-market model and trying to close deals, get validation that people are using and staying with the platform. If you look at how you build a startup, there’s the product-market fit stage, and then the next stage is product-channel fit—what is the effective go-to-market motion to sell this? You start with one channel but then it makes sense to start testing others. Typically, the main models are inbound, outbound, and partnerships, but there might be other routes, like ISV. We monitored four channels: inbound, outbound, and partnerships, and tracked variations like whether inbound is purely self-serve or has sales involvement. Then you start putting small bets on them, monitor how well they perform, and look at metrics like LTV to CAC. It takes time to understand churn across different channels, so in the beginning, the payback time is the best metric to look at—how quickly are we getting the sales investment back? It’s easier to calculate—what are the salaries of the salespeople and marketing costs? After that, you just put more money behind the ones that work, keeping scalability in mind.

The great thing about outbound sales is that there’s no hard limit. You can hire many salespeople, and if the market is big, it will scale. The challenge within marketing, especially with Google Ads, is that it’s limited by search volume. As you put more budget behind Google Ads, Google starts showing it to less relevant audiences. The conversion rates from those added are lower because the averages come down, especially for the new audience, which typically performs poorly. So, you try to find a sales channel that is profitable, scalable, and predictable. If you find that, you can tell investors, “Hey, give us one euro, and we’ll turn it into five euros in 12 months.” That’s how you do it.

You also mentioned that when you added the sales team to the free trial motion, you made them call the most promising leads. You were also keeping track of that investment in sales and marketing to ensure it was profitable. Can you talk more about how you do that, because this is a challenge for many companies? How did you know it was actually additive to results?

We looked at the conversion rates per segment. What we learned was that when we didn’t add a sales touch, the conversion rates were around 10%. After we added a sales touch, the conversion rate increased to about 40%—four times higher. Then we tracked metrics like how many deals one salesperson could handle in a month and how much they could close. This helped us calculate effectiveness. Initially, we only called the biggest trials we got, but we expanded the scope to smaller trials and searched for the best-fit customers. You eventually end up calling about 50% of the largest trials. But it’s important to know when to give up—how long do you keep selling to them? Many companies make the mistake of sticking too long with clients who aren’t ready to buy. What we did was move to the next one quickly. We also had a premium version, so if someone didn’t buy during the trial, they could keep using a limited free version. We didn’t lose them; we just followed up with them after six months when the product had developed a lot, making it more logical to get back to them.

Hey, Would you want to have like a new trial in one year and evaluate again? No one wants to give up, that's a really good insight.

When you think through the journey, were there any things that just plainly failed? What do you remember? Things you tried? I need to think about that. Like, especially in the previous company, going international with the sales SL model failed. That was the learning there. When it comes to Leadfeeder, we could have started to do the other channels earlier. We were quite religious on the inbound model and didn’t really want to add the outbound model on top of it. It was quite late in our journey when we added that. We could have done that way earlier. But there wasn’t anything that really failed. Of course, in the beginning, things don’t work, but if you have the analytics in place, then we were able to make them work at some level, at least. The partnership model is something that we have been struggling with—how do we get the marketing agencies to push us even more forward? So that hasn’t been as effective as we thought. What is the business model for the partner to push us forward? Also, they are many times expecting us to market them to our customer base, which we are not really doing. So, I wouldn’t say that we failed here, but we haven’t really nailed this one at least yet.

MH, so coming down to the people—what were some of the key roles that you had in place to make all of this happen? In the beginning, we only had support. So, we had trials coming in, and we only had support through Intercom. So, we had—and still have—a chat support in the product. So, that’s the main way how customers can get help. After that, we started to recruit salespeople, and we wanted to, we saw that we don’t need these high-velocity salespeople. It’s more people that can help the customer take the value from the product and understand inter-marketing. So, it’s this sales assist role. So, you’re not really selling; you’re just, you know, showing the value of the product, and then the customer just wants to buy. It’s a bit different focus there when it comes to selling. So, we tried to hire people that know broadly about marketing and can add value to the customer. It took quite a long time before we added customer success into the mix. But then we really, really saw that when the product got more complicated, we needed to have people to help customers go beyond the first use case they saw. But we’ve always tried to keep the product as simple as possible so we wouldn’t need to spend too much time there. And because we came from this automated sales, where volume is the key in the beginning, we also made sure that the automated onboarding works well. We had support materials and tried to do things in a scalable way because we already had the volume.

Good, so then last question for today. Thinking back on the whole journey, would you have done anything differently? Probably, like, it’s easy to look from history, but I think the journey we went through is quite logical. I wouldn’t say that there are big things that could have gone differently. We could have maybe scaled this even more aggressively in the beginning, but I think the way we did it makes a lot of sense. We didn’t raise that much money—only a couple of million—and we were able to be quite cost-effective with the investments we got in.

Great, cool. Well, thanks for taking the time to share your experiences today. Is there anything you want to add before we close off? Well, building a startup takes a long time, and especially the early years are many times really, really a struggle. But once you hit the product-market fit and product-channel fit, then it starts to go really well. If you just keep your eye on the customers and the numbers, and I would really emphasize making sure that in the beginning, the KPIs and the product analytics are correct because that was the instrumentation we used to fly the plane, so to say. Otherwise, you are flying blindly if you don’t have that in place.

All right, great. Thank you, Pekka. Thank you, Vincent.

All right, that's it for this episode. Another great conversation today. Clearly, the focus on usage before even charging money for the product helped Leadfeeder build a strong product experience. Beyond that, it's the little things that make a model work, like knowing when to stop chasing leads. Also, be on the lookout for my upcoming book about product-led sales. To get early access, go to plgandsales.com.

Thanks for listening, and see you next time.