How to find the three bottlenecks costing your operation the most

The signal: your team isn't slow. Your system is.

Every UK mid-market operator I talk to opens with some version of the same line: "we're busy, but we're not getting traction." The instinct is to hire. Or to buy another SaaS tool. Or to find a better project manager.

Almost always, none of those is the answer.

The real answer is structural: the same three operational taxes are silently draining the business, and until you name them, you keep paying them.

Tax #1 — The Headcount Tax

What it looks like. Your senior people — the ones you hired for judgement — spend 30% to 60% of their day on work that does not require judgement. Admin. Follow-up. Reformatting. Dictation. Internal status updates. Lookups across half a dozen systems to answer a single question.

What it actually costs. A senior fee-earner on £60k who spends 40% of their week on admin is £24k of salary being spent on clerical output. Across five seniors, that is £120k a year — gone. Worse: you cannot easily reclaim that capacity by hiring junior staff, because the work is too fragmented to delegate cleanly.

How to check it in your own business. Pick a senior person whose work you understand well. Shadow them for half a day (or ask them to log their work in 15-minute increments for a week). Tag each block as either judgement work or clerical work. If clerical exceeds 30%, you are paying the headcount tax at a rate that compounds with every senior hire.

Tax #2 — The SaaS-Sprawl Tax

What it looks like. You are paying for five, six, sometimes eight SaaS tools that each cover 60% of what you need. The other 40% — the bit that makes your operation different — lives in the gaps: a spreadsheet, a private Trello board, a Slack workaround, a manual reconciliation step every Friday.

What it actually costs. Three costs, all hidden: (1) the licence fees themselves — usually £1k to £5k per month for mid-market stacks; (2) the integration effort to keep the tools talking to each other, often a slice of someone's full-time job; (3) the slow erosion of trust in the data, because every system has a slightly different version of the truth.

How to check it in your own business. Count the SaaS tools your team actively uses in the course of a normal week. Then count the workflows that span three or more of them. If you have more than five tools and more than three multi-tool workflows, the sprawl tax is meaningful.

Tax #3 — The Founder-as-Bottleneck Tax

What it looks like. Decisions queue at one person. Not because that person wants control — usually they hate it — but because no-one else has the context, the authority, or the documented criteria to make the call. Deals stall. Quotes wait. Hiring decisions sit. Project starts slip.

What it actually costs. The cost is measured in cycle time: how long a decision takes from triggered to made. A quote that should go out in an hour goes out in two days. A hire that should be approved by Wednesday gets approved by the following Tuesday. Multiply by deal volume and you have the cost.

How to check it in your own business. List the last 10 significant decisions your operation made. For each, note who triggered it, who finally made it, and the cycle time. If a single name is on the "who made it" line more than seven times out of ten, the founder bottleneck is the dominant operational tax.

Which tax matters most for you?

The three taxes do not weight equally for every operation. The diagnostic question is not "which one are we paying?" — usually you are paying all three — but "which one is compounding fastest with growth?"

• The headcount tax compounds linearly with every senior hire. It is the dominant tax when you are scaling people.
• The SaaS-sprawl tax compounds geometrically with every new tool. It is the dominant tax when your stack is mature and growing.
• The founder bottleneck compounds with decision volume. It is the dominant tax when revenue is climbing and the operation has not delegated authority in step.

The decision matrix: build, hire, or change process

Once you have named the dominant tax, you have three honest options:

1. Change the process. Sometimes the bottleneck is the process itself, not the tooling. If the decision criteria can be written down clearly and delegated, that is the cheapest fix. No AI required.
2. Hire. The right answer when the work genuinely needs judgement that cannot be encoded. Senior fee-earners. Specialist engineers. Heads of function. Hiring is the wrong answer when the work is fragmented across many tasks and too dependent on context that lives in one person's head.
3. Build. The right answer when (a) the work is repetitive and rule-based but the rules are specific to your operation, (b) you cannot or do not want to put the data into a SaaS AI tool, and (c) the volume of work justifies the engineering effort. This is where bespoke AI fits.

When AI is the answer (and when it isn't)

Bespoke AI is the right answer when:

• The bottleneck is high volume — the work happens often enough that automating it is worth more than the build cost.
• The work is rule-based but the rules are yours — not what a SaaS template assumes.
• The data is sensitive or regulated enough that you cannot put it through a third-party AI API.
• The bottleneck is load-bearing on growth — fixing it unblocks revenue or capacity, not just convenience.

Bespoke AI is the wrong answer when the bottleneck is genuinely a process problem, a hiring gap, or already adequately solved by an off-the-shelf SaaS tool. The discovery call is where you sort which is which.

Next step: book a discovery call

A 45–60 minute technical conversation. We map your operation, name the three biggest leaks, and tell you which is the right candidate to fix — bespoke AI, process change, or hire. You leave with a costed bottleneck map. No deck. No pitch. No follow-up sequence.