Sales Process Automation Starts Upstream

Most sales automation projects start where the work is visible: at outreach. Sequences get built, templates get stored, cadences get timed. Afterwards the same process runs faster and wider, and delivers the same hit rate across a larger pool.

Manuel Hartmann, Founder and CEO of SalesPlaybook, discussed this with Christoph Burkhardt, who built the Swiss market for financial software vendor LucaNet over twelve years. The conversation is published in his LinkedIn newsletter Diary Of A CRO, and its conclusion is uncomfortable: the leverage sits upstream of the process everyone optimises.

This article walks the four places where that holds, and names the symptom that tells you each one is stuck.

From the LinkedIn newsletter Diary Of A CRO by Manuel Hartmann.

The arithmetic that comes before any tool

Before you spend money on automation, there is a calculation almost nobody runs. Two conditions decide whether a company can buy from you at all, and they multiply rather than add. Skip them and you simply enlarge the pool of addresses where even a flawless conversation changes nothing.

The first condition is the level of decision. If your product solves a task that gets decided at group level, the local subsidiary of a foreign parent is not a customer, because the decision happens elsewhere. The second condition is timing. Accounting or resource planning software gets replaced perhaps once a decade. The probability that a perfectly matched company decides in your category this particular year is therefore a fraction.

Burkhardt calls these two "factor-10 effects", because each one thins the addressable market by an order of magnitude on its own. What matters is that they compound.

Our take: this is why every pipeline engagement we run starts with those two questions rather than with tool selection. Automation pointed at an unfiltered list makes waste cheaper. It does not make it smaller.

How you spot it: your outbound numbers look bad, and the internal answer is "more volume".

Brief your top of funnel for the list

What your top of funnel delivers depends on what you ask it for. Ask for leads and you get a stream: a handful of contacts each week that nobody cares about after the quarter closes. Ask for the list and you get something that stays and compounds.

In Burkhardt's setup, half a headcount was briefed to deliver a fixed number of newly qualified target accounts per week. Qualified meant three questions answered. Is this genuinely our ideal customer profile? Is this the headquarters? And who owns our topic there? Leads arrived as a by-product.

That inverts what most teams do with their top of funnel. The difference only shows over time: a target account list grown deliberately over two or three years pays off at the moment a company on it decides, and you are not the vendor introducing yourself that week.

Our take: we create the target account list as its own object in the CRM before the first campaign runs. In HubSpot that is a maintained company list carrying the three qualification fields, and it becomes the asset every later workflow reads from.

How you spot it: you cannot say in one sentence how many qualified target accounts your list holds today, and how many were added last week.

What you can automate, and what you cannot

This is where automation either returns capacity or makes your outreach generic. Teams usually automate where it is technically easiest, which is outreach, and leave list building as manual work. That split is the expensive one: you scale the replaceable part and leave the valuable part untouched.

The useful line runs between volume and depth. Building and enriching a target account list is rule work: fixed criteria, repeatable checks, many similar cases. That scales by machine. Burkhardt describes how one person with current tooling covers a multiple of the accounts per week that were feasible by hand a decade ago.

In our delivery, Clay handles that step: enriching firmographics, resolving headquarters and group structure, finding who owns the topic. What comes out lands in HubSpot as an asset and becomes measurable there. The letter, the event and the conversation with the decision maker stay human.

The reason is structural. A language model simulates intelligence, but when something is missing from its training data it produces an answer anyway. A person who does not intend to deceive will abstract instead and still answer correctly. For your revenue organisation that means machines are strong at volume and consistency, and weak exactly where trust and domain depth decide.

Precision has become more valuable because of this, not less. Once everyone can send faster and cheaper, precision is the remaining difference.

How you spot it: your first touch is generated and your list is hand-maintained.

Domain depth beats methodology

The largest measured gap between two sales people is rarely selling talent. Value selling is undisputed as an ambition, but it often gets read as vocabulary: five benefit statements, well phrased, and the rest will follow. A decision maker can tell within minutes whether somebody understands their problem or is paraphrasing it.

Burkhardt describes a case from his own team in the newsletter: a consultant finally got a shot at a sales role and reached a markedly higher win rate from a standing start than expensively hired enterprise profiles with well-known logos on their CV. Same product, same market, same price book. The difference was domain depth.

His response was not to hire only consultants. It was more expensive than that: more than half his leadership time went into coaching, and every lost deal was reviewed together. Depth can be hired or built. Skipping it is the one option that does not work.

We saw the same lever in a different place. For Jua, the decisive change was the level of the target persona: selling to decision makers instead of analysts within less than 4 months. Same mechanism as the first condition above — work past the decision maker and you hold good conversations with no outcome.

How you spot it: your best win rates sit with individual people and nobody can explain why.

Make the ritual do the policing

Clean CRM data gets demanded in most teams and rarely produced. That costs twice: leadership spends its time chasing instead of coaching, and the numbers it looks at are wrong anyway. One habit fixes both, and it costs no extra time.

Burkhardt opened the same dashboard first in every weekly one-to-one. Every week, same order. The team knew it and cleaned up beforehand. Quick look, all green, on to the deals. No memo, no chasing.

The mechanism is the incentive rather than the control. The question shifts from "will my manager complain" to "do I want to be the red row". People respond to that, data hygiene included.

Behind it sits a second observation that reaches further. A leader without deep product and market knowledge is left with controlling individual fields as the only available instrument. Micromanagement is therefore a diagnosis, not a leadership style. When you see it in your organisation, the right question is not "how do we lead better" but "what does this person not understand".

How you spot it: your pipeline reviews spend more time on the state of the data than on the deals.

Focus on whether, not when

Deals are rarely lost these days; they die slowly. Q1 is too early, Q2 is full of conferences, Q3 is holiday season, Q4 is annual planning. The rep will not pull the deal from the pipeline because a forecast depends on it. The buyer will not say a final no. The result is a pipeline that looks maintained and carries nothing.

The answer is unspectacular: Burkhardt stopped negotiating the date. The focus was on whether the deal closes, not when. If it fits six months from now, a follow-up date six months out is the answer. Done.

That is the arithmetic from the top of this article read backwards. If your buyer decides once a decade, the quarter boundary is an invention of your reporting rather than a feature of the market. A large share of lost deals carries the reason "postponed", and that is a follow-up date.

In practice the direct question beats pressure: "My sense is this is not on your plate right now — should we look again in a year or two?" That usually gets an immediate answer, and it gives you capacity back.

How you spot it: your forecast slips by the same deals every quarter.

The four levers at a glance

Each of the four places shares one structure: it sits ahead of the process you would optimise. Automation amplifies everything downstream of it and leaves everything upstream untouched.

Where to start this week

Three questions you can answer before you pick a tool.

First: how many companies on your current list can actually decide in-house? Count headquarters rather than addresses. Second: how many qualified target accounts were added in the last four weeks, and who owns that brief? Third: what share of your lost deals carries the reason "postponed", and does every one of them sit on a follow-up date?

If any of the three leaves you without a number, you have your answer about where automation should start.