A free assessment for small businesses
Find out how many hours a week you would get back, and what to start on Monday.
This is the interview consultants charge around a thousand euros for, except you run it yourself and it takes about fifteen minutes. At the end you get a number in hours, a four day plan at ten minutes a day, and a list of what you must not automate yet.
28 questions. No sign-up, no email. Answers never leave your browser unless you choose to email them.
28
questions, one per screen
Four about the business itself, 24 about the work, the people and the risk. One fewer if you work alone.
4
day plan, ten minutes each
A report nobody can act on is not a report, it is an excuse.
0
answers leave your computer
Your answers are never sent to a server. The scoring runs in your browser and the answers stay in this tab.
The assessment
28 questions about your business
Answer as things are, not as you would like them to be. A bad answer is more useful here than a good one: it is the one that shows where the hours are.
1 of 28. How many people work in the business, including you?
How it works
Four steps, fifteen minutes
Context
Your size, your trade, and the one thing you would delete from your week. So the advice is your size, and it opens with what you came for.
The interview
24 questions, one screen at a time. Where the hours go, where the data sits, what happens to the odd cases, and what the team would say.
The arithmetic
At the end it asks what an hour is worth, and turns the hours into money a month. The working is shown, not just the answer. Skip that question and you get hours only, rather than a figure nobody chose.
The plan
Quick wins with their cost and setup time, a four day start, the larger projects for later, and the brakes you must not drive through.
Why this looks different
Most assessments ask about technology. This one asks about your work.
The documented process is almost never the real process. Asked how it starts, people say an email arrives. In fact it arrives from forty senders, no two alike, sometimes a PDF, sometimes a photo, sometimes a forwarded thread, and half the cases are exceptions whose handling rules live in one person's head. That is why this asks about exceptions and not only about systems.
And that is why it asks about people. Rollouts do not fail on technology. They fail when the team believes this is about their jobs, at which point output gets checked twice, mistakes go unmentioned, and the new process gets worked around. From outside it looks like poor technology.
Every recommendation has to pull one of three levers or it does not belong: make more money, save time, or improve what the customer receives. If it pulls none of them it is not a solution, it is just another tool.
Working together
Three steps, starting with one week
The assessment above is free and finished: nothing in it is blurred out and there is nothing to unlock. If afterwards you want somebody to do this with you, here is what I offer and what it costs.
EUR 900, one week
The analysis
The same thing you just did, but with me and on your real data: your two most expensive processes written out step by step with their exceptions, and one working prototype built. About half of firms should not proceed after this, and when that is the answer I say so.
from EUR 3,000, fixed
The build
A fixed price for an agreed outcome, not for hours. Acceptance criteria are written down before work starts, so there is nothing to argue about at the end. It is built on top of what you already run rather than by moving you onto something new.
from EUR 900 per month
Alongside
Two sessions a month in which your own people build the next thing while I sit alongside. This solves the item almost everyone fails on: that somebody inside the firm can keep it working once I am gone.
No commitment attached. If waiting is the right answer in your case, I will say that too. If someone else makes the decision, bring them to the call.
About
Why small businesses specifically
Because here the distance between a decision and a result is short: one person, or one conversation between a few. There is no committee, and no pilot that quietly ends after a quarter. So a week of work shows up in the same month.
And because this is where most of the damage is currently done. A large company has somebody to check whether the new assistant will show the payroll sheet to everyone. A twelve person firm does not, which is why the list of what not to do yet is not a courtesy, it is the main part of this page.
Common questions
What people usually ask
Is there really nothing to sign up for?
Really. Nothing you answer is sent to a server: your answers are scored inside your browser and kept only in that tab, so a reload does not lose them. Start again clears them. That also means I cannot see who took it. At the bottom of the report there are buttons that open your mail app, and you decide whether to send anything, and to whom.
Where does the number of hours come from?
From your own answers. Three questions ask you to estimate how many hours a week go into correspondence, re-keying and searching. Each of those is multiplied by the share it is realistic to recover, and the report shows every multiplication separately. Those shares are an estimate rather than a measurement, and the page says so plainly rather than in small print.
Why does it ask what an hour costs?
Because without it hours cannot be turned into money, and guessing would mean showing you a convincing-looking figure I made up. You can skip the question. The report then shows hours only, and says why the money figure is absent.
What does it mean that something must not be automated?
Some answers do not lower a score, they stop the advice. If shared drive permissions have never been checked, for instance, a search assistant will find the payroll sheet in a second and show it to whoever asked. Or if half the cases are exceptions whose rules are written down nowhere, there is simply nothing to automate. In those cases a brake appears at the top of the report, with the condition under which it stops applying.
Would our data leave the firm if we worked together?
It depends on the design, and that is decided during the analysis rather than after it. Some work runs entirely inside your own infrastructure. Where an external model is used it is named in the contract as a sub-processor, and the data processing agreement is signed before work starts, not afterwards.
How is this different from a Zapier scenario or similar?
A scenario moves data from one place to another and nothing more. It cannot see what came of it, so a year later it behaves identically. The difference is not the tool, it is whether the result returns and changes the next decision. Often a scenario is in fact the right answer, and when it is, the report says so.