A competent firm should be able to describe, precisely and without defensiveness, the customer who should not buy from it. One sign of seriousness is the willingness to say “you are too early” or “that isn’t our problem” with enough frequency that the word means something. This is an attempt to say it in a structured way.
Operations automation — which is roughly the practice of designing how a business’s repeating work happens, and using software to do more of it without human intervention — pays back handsomely for some businesses and fails expensively for others. The difference is not usually about the business’s size, sector, or founders. It is about a handful of structural conditions that can be assessed relatively quickly.
When automation pays back
A business is ready for this work when four conditions are present at once.
The processes repeat. If the work you would be automating happens three or more times a week, reliably, in roughly the same shape each time — lead intake, onboarding, invoicing, reporting, follow-ups — automation will pay back. If it happens twice a month in a shape that changes each time, the payback is weaker and the maintenance cost gets worse.
You have a stable definition of what “done” means. If the people doing the work can agree on when a case is complete — “the client has received the deliverable and confirmed receipt”, “the invoice has been raised and recorded” — automation can check against that definition. If there is genuine disagreement about when a thing is done (which is more common than people admit, particularly in consulting), automation tends to encode one side of that disagreement and surface the tension.
There is at least one person who can own the operation. The automations we build are not unattended systems. They need a human owner — someone who reviews exceptions, adjusts rules, notices when something new has started going wrong. This does not need to be a full-time role; in a small business it is often a few hours a week for an operations manager, an executive assistant, or a technically-confident team member. But it needs to exist.
You can articulate why the process exists. When we interview your team, we want to hear a coherent answer to “why do we do this at all?” for each process we are about to automate. “Because we always have” is a warning sign, not a diagnosis. Automating a process you cannot justify will produce a faster version of the wrong thing.
When all four of these are present, our work routinely delivers payback within three to six months of the setup phase ending — measured in time recovered for the team, or cost saved, or revenue protected (in the case of lead intake or billing).
When automation does not pay back
The reverse conditions are where we have learned to be cautious.
You are too small
If your business is one or two people and you are doing everything yourself, automation is rarely the right spend. Not because it wouldn’t work — it often would — but because the setup cost is large relative to the amount of time you would recover, and because very small businesses often genuinely benefit from the owner doing the operational work themselves. Handling your own invoicing, chasers, and onboarding teaches you where the weaknesses in your business are in a way that a dashboard does not.
Our loose rule: below a team of about four people, or below about £150,000 in annual revenue, the setup spend is usually better deferred. There are exceptions — a one-person business with very high transaction volume, for example — but not many.
You are pre-product-market-fit
If you are still working out what you sell, to whom, at what price, and how the work is actually done, automating the operation is premature. You will be codifying a process that is about to change. The output of such an engagement is a set of carefully-built automations that are obsolete within a quarter.
Signals you are pre-product-market-fit and should wait: the pitch you use with new prospects has changed substantively in the last six months; your pricing is still being revised; there is internal debate about which segments to target; the team is still arguing about what “good” looks like in the deliverable.
This is not a critical observation. It is a fine place to be; it is just not a good place from which to automate. Get the product-market fit first. Automate the successful version of the operation, not the evolving one.
You do not have someone who can own the system
If nobody in the business can spare a few hours a week to act as operational owner of the automations once we’ve built them, the work will degrade. We design for minimal maintenance, but some is irreducible — a client changes a system, an API deprecates a field, a new process emerges. Someone has to notice and adjust.
We have occasionally built automations for businesses where the owner-founder intended to be that person themselves. It rarely works. Founders are the most overcommitted people in any small business; the operational owner is almost always someone slightly below them in the hierarchy, with the time and the authority to act.
Your problem is not really an operations problem
A surprising number of the enquiries we receive are about problems that look operational but are actually strategic. “Our sales process is inefficient” might mean “our sales process is poorly designed”, which is an operations problem — or it might mean “we are not sure who we are selling to or why we win”. The second is not an operations problem. It is a strategy problem, and pointing a workflow tool at it will not help.
Similarly for some common false positives:
- “Our team is overwhelmed.” Sometimes an operational problem; sometimes a hiring problem; sometimes a pricing problem (you have too many clients because you are too cheap).
- “Our clients are confused.” Often a messaging problem, not an automation problem. A better-worded initial email solves more confusion than any workflow.
- “We can’t see the numbers.” Sometimes a reporting problem; sometimes a data-quality problem upstream; sometimes a consensus problem about what to measure.
Part of our diagnosis is distinguishing these. We have, more than once, told a prospective client that the problem they described was not a problem we could help with and that they should spend the money elsewhere.
You are in crisis
If you are writing to an operations firm on a Wednesday afternoon because something is going badly wrong by Friday, you are probably not the right fit for us. Good operational work is slow, reflective, and cumulative. It rewards a steady engagement at normal pressure. A firm responding to a crisis — a regulatory inquiry, a key-person departure mid-engagement, a client threatening to leave — needs a different kind of help, often delivered faster and rougher. We are not the right shape for it.
Occasionally we will take on a crisis engagement for an existing client, because we know the territory. We rarely take one from a stranger. If you are in crisis and reach out to us, we will usually suggest some people who are better suited, or recommend waiting until the fire is down to embers before we talk.
A ten-minute self-audit
If you would like to check your own readiness before writing to us, here is a quick version. Answer the following with an honest yes or no.
- Is there a repeating process in your business that happens at least three times a week in roughly the same shape?
- Can the people doing that process agree on when a case is “done”?
- Is there at least one person in the business (not necessarily you) who could own the maintenance of an automated version — a few hours a month?
- Can you articulate, in a sentence, why the process exists?
- Has your pricing been stable for at least six months?
- Have you been in business for at least a year?
- Is your team four or more people, or is your annual revenue above £150,000?
- Are you not, at this exact moment, in a business-threatening crisis?
Six or more “yes” answers, and you are probably ready. Three to five, and the answer is “probably not yet, but let’s talk about when you will be”. Two or fewer, and any firm telling you that you are ready is trying to sell you something.
A note on the counter-case
None of this is absolute. We have been wrong. We once declined an engagement from a business we assessed as pre-product-market-fit; they went and hired a different firm, had the work done, and made the automations work by sheer force of founder energy. We were wrong about that specific business.
The heuristics are good over many cases, not good over every case. If you think your business is an exception to one of the above — a one-person business with unusual volume, a pre-PMF startup with a stable repeating sub-process, an owner who genuinely has the time to own the ops — we’d be interested to hear why. We are more interested in being right than in being consistent.
How to begin, if you think you’re ready
Book a discovery call. Thirty minutes, free. If you are not ready, we will say so, and sometimes tell you what to do in the meantime. If you are, we will send a diagnosis within the working week and you can decide from there.