Your product is excellent. Your sales aren't accelerating. The problem isn't what you think.

Nine out of ten tech founders I meet have the exact same problem.

Their product is excellent. The demo wows. Prospects are sold within thirty minutes.

And six months later, the deal still isn’t signed — or the rollout still isn’t finished.

This isn’t an isolated case. It’s a pattern.

I see it in Series A-funded companies, in post-Series B scale-ups, in Belgian tech SMEs doing €3M who want to push to €10M. The sectors change, the tech changes, but the shape is always the same.

On one side, a real innovation — often impressive. On the other, a commercial cycle combining heavy installations, complex integrations, legal or regulatory validations, and a user experience that has to be simplified to the bone for the end client to actually adopt it. In between, exhausted teams chasing deals that won’t move.

And almost always, the founder says the same sentence: “We need to automate more.”

At that stage, it’s usually the worst thing you can do.

The trap of premature automation

Tech founders are logical. They have an excellent technical product, so they want an excellent technical sales process. They buy a new CRM, plug in a sequencing tool, hire a Head of Revenue Operations, and put an AI agent on inbound qualification.

Automation feels like progress. Dashboards fill up. KPIs exist. You can pitch it to the board.

Except at this stage, the problem isn’t a lack of tools. It’s a lack of understanding.

In a complex B2B sale, what you’re actually trying to do at the start isn’t to sell more. It’s to understand precisely what you’re selling, to whom, why, and with what real frictions along the way. That understanding cannot be automated — because it doesn’t exist yet.

You can’t automate something you haven’t first industrialized by hand.

Do things that don’t scale — what it actually means

Paul Graham wrote that piece in 2013. Every founder I meet has read it. Almost none actually applies it.

Because they interpret it as “do artisanal things at the beginning, it’s cute, it’s a phase.” That’s not what it is.

Do things that don’t scale, in a B2B tech company, means: the founder runs the sales themselves until they have 10 signed deals that follow the same pattern. The CTO does the installations personally until they see what breaks systematically. The Head of Customer Success follows the first 20 accounts by hand, not through a tool.

Not because it’s romantic. Because that’s where the real information lives. The kind of information you will never pull from a Gong dashboard, a Salesforce report, or an AI agent listening to your calls.

Until you’ve done that, you haven’t earned the right to automate. And if you automate anyway, you automate the wrong process — and you spend the next 12 months undoing it.

What we learned at Sortlist

At Sortlist, one of the hardest parts has always been the matching between demand and supply, around a concrete project.

A company looking for an agency has a fuzzy need. An agency selling services has a broad promise. In between, there’s a matching workflow: understanding the brief, qualifying the budget, selecting the five right agencies out of twenty thousand, introducing them properly, following through to deal signature.

For years, we tried to automate this aggressively. Algorithms, scoring, tags, automated matching, rule-based routing. Every attempt solved 70% of the problem and added 30% of friction. And buyer NPS refused to climb.

The truth we took too long to accept: on the part of the process that really mattered — the quality of matching — no algorithm we had at the time could beat a human who had spent 90 minutes on the brief.

We ended up keeping a team of matchmakers doing that work by hand, at scale, while the tech handled pre-selection and industrialized the rest. We didn’t automate the core. We industrialized a manual process.

Deals started closing faster because they were better qualified upstream. Supply churn dropped because agencies received leads that were genuinely relevant. And from there — only from there — we could start automating specific pieces.

The real question to ask

The question isn’t: “how do I automate my sales process?”

The question is: “where, in my value chain, am I losing the most time and quality?”

And once you’ve identified it: “can this be automated without killing what makes it work?”

The answer is often: not yet. And that’s fine. What you need at this stage isn’t to automate — it’s to understand that bottleneck well enough that you can attack it with the right intent, later.

The blue book from 1984 that settles the debate

There’s a book I recommend to almost every tech founder I meet. It was written in 1984. The cover is blue. It’s called The Goal, by Eliyahu Goldratt.

It’s not a business book in the classic sense. It’s a novel — the story of a plant manager trying to save his factory before it gets shut down. Behind the fiction, Goldratt lays out what he calls the Theory of Constraints.

The central idea: in any production chain, there is always ONE bottleneck that determines the throughput of the whole system. You can optimize every other step as much as you want — if you don’t attack the bottleneck, you produce exactly the same amount.

Applied to a B2B tech company, this is devastating.

Because most founders try to optimize everywhere at once. They invest in marketing, in the CRM, in sales hiring, in product, in brand. And they’re surprised ARR isn’t climbing.

In reality, there’s probably ONE specific step — the demo, the technical implementation, the legal validation, the pricing negotiation, the integration into the client’s IT stack — that is the bottleneck. Until it’s fixed, nothing else matters. Hiring three more SDRs while your integration cycle takes four months will only stack up a pipeline that doesn’t convert.

How to actually map this

A simple exercise, half a day with your two or three key people:

One — Draw the complete client workflow, from first contact to the moment they actually generate value with your product. Not the marketing version. The real version, with all the back-and-forth.

Two — At each step, note the average time spent on the client side, the average time spent on the team side, and who does what.

Three — Identify the step that creates the longest calendar delay between two milestones visible in the client experience. It’s not always the step that’s heaviest in hours. It’s the one where the project “waits” the longest.

Four — That’s where you attack. First. Before anything else.

Most founders are surprised when they do this exercise. It’s almost never the step they thought. It’s often further upstream — in qualification, in the delay between the verbal “yes” and the start of the POC, in the onboarding of a specific persona on the client side (legal, security, IT) that you’d forgotten about.

And very often, that bottleneck doesn’t get fixed with a tool. It gets fixed by changing who does what, when, and with what authority.

What I’d tell a founder who wants to accelerate now

Three simple rules, which I try to install in the companies I work with.

One — No automation before 10 signed deals following the same pattern. As long as the pattern isn’t stable, there’s nothing to automate. You’d be automating noise.

Two — Every quarter, founders spend at least three full days shadowing the sales and delivery process end-to-end with a real client. Not reading reports. Not listening to recorded calls. Sitting in meetings, reading the emails, watching where the client actually struggles.

Three — Before scaling anything, map the value chain, identify the bottleneck, and fix that — not everything else. A good scale is a scale that attacks the constraint. A bad scale is a scale that stacks capacity on steps that weren’t the problem to begin with.

What I’d do differently at Sortlist

For too long, we believed that a large-scale marketplace had to run exclusively on tech. We delayed industrializing the manual work because it felt less “scalable,” less “SaaS,” less “valuable” to a fund.

If I were building Sortlist today, I’d start with ten perfect manual matchings. Zero algorithm. Zero tool. Just a human, a phone, and twenty minutes of briefing with each side.

And I wouldn’t write a single line of code to automate anything until I’d understood precisely what has to stay human, what can be assisted, and what can actually disappear into an automated workflow.

That alone would probably have saved eighteen months of engineering and three product pivots.

The short conclusion

Accelerating sales in a B2B tech company isn’t a tooling question. It’s a question of where you invest your attention.

And the discovery most founders hate making is this: the answer is almost never in the CRM. It’s almost always in a step you don’t want to look at — because it’s manual, painful, and feels like a weakness.

That manual step is your leverage. Map it. Attack it. Everything else follows.

Thibaut Vanderhofstadt is co-founder of Sortlist — Europe’s leading B2B matchmaking platform, active in over 140 countries. He now works with SaaS founders, marketplace operators, and ambitious non-tech companies through MetSaaS. Book a diagnostic →