Your onboarding is a pressure test

If I get one process to examine in a company, I ask for onboarding.

Not because it's the most important — retention usually matters more, and product quality more than that. Because onboarding is the process that crosses the most seams. A new customer touches sales, provisioning, data migration, billing, support, and product in their first two weeks. Every one of those transitions is a joint, and every joint is a place your pipe either flows or doesn't.

Onboarding isn't just a process. It's a pressure test that your customers run on your entire operation, for free, continuously.

What the readings tell you

Time from payment to first real use. Not first login — first time the customer did the thing they bought the product to do. This number is the closest thing to a single-figure summary of your operational health, because it can only be short if every segment and every seam works.

Every day in that number is sitting in a specific place. Find it. It's almost never where the team assumes; when we measured it at one portfolio company, 60% of the elapsed time was a single seam where provisioning waited on a data file nobody had asked the customer for yet.

Number of times the customer has to repeat themselves. Count how many times a new customer supplies information they already gave you. Each repetition is a seam where the upstream Out didn't match the downstream In — a fitting that doesn't fit, exposed by the one person in the process who can see across all your internal boundaries.

Customers experience this as sloppiness. Diagnostically it's precise: each repetition names a specific broken handoff.

Where they go quiet. Instrument the drop-offs. The step where new customers stop responding is the step that's too hard, too unclear, or too far from the value they were promised. Silence is data, and it's the data most companies discard as "they weren't a good fit."

How many humans they meet. Not inherently bad — high-touch is a real strategy. But if a customer meets six people and each one re-establishes context, you don't have high-touch service, you have an uncoordinated pipe with a friendly interface on top.

Why this beats an internal audit

You could map these seams internally. You should. But an internal review has a blind spot: everyone evaluates the process from inside their own segment, where it looks fine, because from inside a segment it always does.

A new customer has no such loyalty and no such context. They experience the whole pipe as one thing, and they notice every joint — because a joint that leaks is precisely what "this feels disorganized" means from the outside.

That's an outside-in view of your operations that you cannot generate any other way. Most companies collect it and then aggregate it into a satisfaction score, which throws away all the diagnostic information and keeps the part that makes people feel bad.

The instrumentation

Three gauges, and they're cheap:

Median and P90 days from payment to first real use, with the elapsed time broken down by step so you can see where it sits.
Repeat-information count — how many times the customer supplied something you already had.
Step-level drop-off — where do new customers go silent.

Publish them weekly. Give each one an owner. And review the P90 case specifically, by name, every week — not the aggregate. The customer who took thirty-one days had a specific experience with a specific cause, and that cause is present at some rate in every onboarding you run.

The thing to resist

The reflex when onboarding is slow is to assign someone to shepherd customers through it. A dedicated onboarding manager. A concierge.

That works, immediately and visibly, which is why it's so tempting. It also hides the problem permanently. You've hired a person whose full-time job is compensating for seams that don't fit, and once that person exists, the seams will never be fixed — because from the outside everything now looks fine, and the cost has been converted into a salary line that grows with your customer count.

Fix the fittings first. Then, if high-touch is genuinely your strategy, add the human — and let them do work that adds value instead of work that patches leaks.