In this conversation, Jonathan Wuurman from Luzmo talks through how the team turned a CRM that felt like a data graveyard into a repeatable outbound system for finding the right accounts, the right personas, and getting to sales-ready conversations much faster.

We also get into why inbound started changing for them, why more data was not the answer, how they pressure-tested Evergrowth with a focused proof of concept, and what they learned about ICP, segmentation, and building an outbound system that improves over time.

Looking back, what was the main challenge you were trying to solve before Evergrowth?

Jonathan:
If you take Luzmo, we were in a luxury position, honestly. Around 95% of the business was built inbound. That is a beautiful place to be. It was a healthy flywheel. The company had been in the market for ten years, and we had strong content and SEO working for us.

But we started to see volumes come down a bit. I think part of that is simply that ten years ago, the market was less aware of what we offered, and today it is much more saturated.

At the same time, our ACV changed. Luzmo started as something much more product-led, with a very different kind of deal motion. Today, our average ACV is probably somewhere around 25K to 30K a year, so it is a different kind of sale.

So at some point, you need to inject more sales-led motion into the business. And we did not really have that yet. The AEs were used to getting inbound. The outbound engine was not there in a real way.

On why “more data” was not the answer

Jonathan:
What I saw in the market was the usual playbook: you buy a big database, you build your ICP and personas in there, it spits out a list, and then you start mass emailing and mass calling and hope that by throwing enough things at the wall, something sticks.

I just do not believe that works.

That might have worked in a different time, when email response rates were much higher and the channels were less crowded. But today, if you are reaching out to hundreds of companies and multiple people in each one, and you are expecting high reply rates, I would love to know the recipe.

That is where I started thinking differently. I thought, maybe agents can help me do what I already see good people on my team doing manually.

Because what you want to give to an agent today is the repetitive work — the tasks a robot can potentially do as well as a human — so your people can focus on the value-added part.

Can you give an example of that kind of work?

Jonathan:
Yes. A very simple one.

When one of my sales people gets an inbound lead, one of the first things they do is go to the website. They look for signals. Is there a login button? Is there a trial? What is in the navigation? Is there anything around product analytics? Better decision-making? Those kinds of things.

That is not just random browsing. That is experience. That is pattern recognition.

So when we built our ICP logic, I wanted to ship that way of thinking into the process. I did not want ICP to be only firmographics on a spreadsheet. I wanted the Evergrowth agent to type in the website, visit it, and do the same kind of first-pass research that one of my team members would normally do.

That is where I saw huge value. If the agents can handle that part, then my people can spend more time building the relationship, using the context, understanding intent, and doing the human work.

On the “CRM data graveyard” problem

Jonathan:
The first use case for us was very simple. I am convinced that CRM systems, or data lakes, or warehouses — whatever you call them — often become data graveyards.

You run a business for years. You collect inbound leads, event lists, ebook downloads, contact imports, all sorts of things. You connect tools, you add more and more data, and at some point you become a bit like the Cookie Monster. More, more, more.

But the real question is: what are you actually doing with that data?

We had just hired what I call an OG — a guy whose superpower is calling. And I genuinely think that is a superpower. Picking up the phone and calling people who did not ask to be called is hard. Most people hate it. It takes confidence.

So I looked at the CRM and said: let us use this properly.

We pulled out 89 companies that had touched Luzmo in some way over the last ten years. Maybe they downloaded an ebook, maybe we met them at a trade show, maybe they came in through some other source. They were in the CRM.

Now the question became: which of these are actually worth working?

How did you approach that with Evergrowth?

Jonathan:
First, we used an ICP agent.

For me, ICP can become either way too simple or way too complex. I try to sit somewhere in the middle. For example, B2B SaaS is still very broad. So instead of just looking at company size, I care about things like the size of the product team, because that is more relevant to whether Luzmo is a fit.

We ended up with 18 ICP attributes that we wanted to validate.

Now imagine doing that manually across a list. A human could easily spend 30 minutes, maybe even an hour, per company, finding and checking all that information.

So we built that logic with Evergrowth. Out of the 89 companies, Evergrowth validated 37 as real ICP matches.

That already told us something important: if I had reached out to all 89, a lot of that would have been bad marketing. I cannot sell to companies that are not the right fit.

And once you had the right companies, what happened next?

Jonathan:
That is where personas come in.

At the company level, you are still very high-level. At the end of the day, you do not do business with companies. You do business with people.

So we took those 37 companies and looked at the contacts we already had attached to them in the CRM. There were 278 contacts.

And out of those 278, only 4 were still valid personas.

That was the big wake-up call.

People move jobs all the time. Contacts go stale all the time. So for me, that proved the point: the freshness of your CRM data should be goal number one, not just having more data.

Accurate data matters much more than having loads of data that is no longer useful.

On rebuilding the buying group

Jonathan:
At that point, we had 37 valid companies and only 4 valid contacts. That is obviously not enough to build pipeline.

So I went back and said: okay, now that we have an agent to validate ICP and another one to validate persona, can we build an agent that actually finds those personas inside the right companies?

That is how we got started on the enrichment side.

And from there, we rebuilt the list to 73 relevant contacts across those 37 companies. So on average, around two right people per company.

That changed the game, because now the caller was not working blind. He had the right type of company, and the right type of person inside it.

On qualification and readiness, not just meeting booking

Jonathan:
One thing I feel strongly about is that SDRs were originally meant to book meetings. But the reality is more complicated than that.

I do not just want somebody throwing meetings on an AE’s calendar. I want someone assessing whether the account is actually worth working, and where it sits.

So in our process, George is mainly doing one thing: he is putting accounts into buckets.

It is basically things like:

• pain not aware, solution not aware
• pain aware, but not solution aware
• pain aware, solution aware, but not on the roadmap
• pain aware, solution aware, and on the roadmap

That is the key thing I ask him to do.

I do not want him pitching. I do not want him forcing demos. I want him to understand where the account is and give us a proper starting point.

Then the next step is for someone else to take that context and build the relationship from there.

What kind of results did that produce?

Jonathan:
I can share the Netherlands example.

We had 317 contacts in that process. After working through ICP, persona, and enrichment, George managed to reach 239 of them in the sense that he at least left a message and sent a follow-up email. So that was what I considered valid data in execution.

That is also where Evergrowth beat one of the big US vendors in this space. I will not name them here, but everyone knows who I mean. We saw around 30% more phone numbers, and according to the AE working the process, around 60% better accuracy. Those are rough directional numbers, but that was the kind of gap.

From those 239 contacts, we ended up with 92 qualified accounts across the buckets.

And from there, we already had 6 reach SQL stage.

Was that a typical speed from outbound to the AE stage?

Jonathan:
No. Much faster.

And I think that is one of the most important things I learned in this project.

The speed did not come from doing more activity for the sake of it. It came from doing the prep work properly, so that when the reps engaged, they were engaging in the right place, with the right people, and with more confidence.

On why better data improves rep confidence

Jonathan:
Calling is hard. It is one of the hardest jobs there is, in my opinion. It is uncomfortable. People hang up. A lot of reps are shaky doing it.

So if I give George a good reason to call you, I make his life easier. I give him more confidence. I also bring more value into the conversation.

There is a huge difference between saying, “Hey, we saw something creepy on your website,” and saying, “Hey, we work with companies like yours that are dealing with this kind of problem, and I think this may resonate.”

That second approach is much more human. It gives the rep something real to work with.

So yes, data is the means to an end. The end is better conversations.

On how Evergrowth helped refine Luzmo’s ICP

Jonathan:
One thing I did not fully understand at the start was how much the agents could help me improve my own thinking.

I remember doing something a bit dirty, to be honest. I pulled 100 companies from our own CRM and gave them to Mert and said, let the agents validate ICP.

At first the number that came back was much lower than I expected. But that is when my brain started to click.

The important part was not just the number. It was the explanation. The agents were telling us why yes, why no, why maybe, why “I do not know.”

And when you can see the reasoning, you can retrain the system.

That became the aha moment for me. We started re-feeding inputs into the agents almost weekly. We kept tweaking and refining.

Eventually, that same type of list got much closer to what I felt was acceptable, and once we were around that 80% confidence level, I was happy to roll it out more broadly.

On discovering new segments

Jonathan:
That reverse-engineering process helped us in two ways.

First, it helped us sharpen our ICP based on our actual customers.

But second, it also showed us that we had different segments inside what we had been treating as one big group.

For example, my core is very much B2B software. But we also have agencies and partners who can redistribute the platform. Those are not the same thing. They have different pains, different context, different messaging.

And then there was telecom.

One of our co-founders is very interested in moving more upmarket in that space — the Proximus, Vodafone, T-Mobile type of companies. And when we looked back into the CRM, it turned out a meaningful share of our historical data was linked to that sector.

That was interesting, because it showed us that the opportunity had been sitting there already. We just were not really segmenting for it properly.

So this whole process did not just improve qualification. It helped us rethink where to go next.