How to delegate effectively as a firm owner

Delegation is a design decision, not a personality trait

Every firm owner has been told to delegate more, usually by someone who did not have to sign the accounts. The advice is correct and almost useless, because it treats delegation as a matter of willingness. Owners who hold on to work are rarely control freaks. They are people who have learned, from experience, that handing something over costs them a fortnight of explaining, produces a version that needs redoing, and lands back on their desk anyway. Having done that three or four times, keeping it becomes the rational choice on any given Tuesday, and the firm quietly organises itself around the owner remaining the bottleneck.

What makes delegation work is not resolve. It is design: choosing the right work to move, in the right order, with an explicit level of authority attached and a written standard the result has to meet. Handing over a task without those three things is not delegation, it is transfer of anxiety — the work leaves your desk, the responsibility does not, and you now have the extra job of wondering how it is going.

There is a structural point underneath this too. Delegation moves individual tasks. It does not, on its own, create the layer of people whose job it is to run parts of the firm; that is a separate piece of work described in how to build a management layer in your firm. But delegation is where it starts, because a manager with no delegated authority is a title rather than a role. And done in sequence, it is the most direct route to the outcome set out in how to make your firm less founder-dependent.

Start with two weeks of honest time, not a job description

Most delegation attempts start with the owner thinking hard about what they could hand over. That produces a list of things they dislike, which is not the same as a list of things that cost the firm. The reliable starting point is a log: two weeks, every hour, categorised at the end of each day rather than reconstructed on a Friday afternoon. Two weeks is long enough to catch a month-end and short enough that you will actually finish it.

The table below is an illustrative log for the owner of a twelve-person firm with about £1.1m of annual recurring fees. It is invented to show the shape of the exercise, not client data. The shape is what matters: about half the week is work that only feels like it needs the owner.

Three things usually fall out of a log like this. The first is that the smallest line is the one the firm most needs — an hour a week of planning, sitting underneath eleven hours of reviewing other people’s work. The second is that “interruptions” is not a category to delegate but a measure of how unclear everything else is; it shrinks on its own once decision rights are explicit. The third is that the two largest delegable blocks — review and inbound client queries — are precisely the two owners are most reluctant to touch, because both feel like quality and relationship rather than process.

Then sequence what you found. Hand over the administrative and approval work first: it is low-risk, it is quick, and it gives both sides an early win that makes the harder handovers credible. Move to review and sign-off second, under a threshold rule rather than all at once. Quoting for standard compliance work comes third, once there is a price list somebody else can apply. Keep the judgement work, the difficult client conversations, the hiring decisions and the direction of the firm. That order matters more than the speed: a firm that delegates review before it has fixed its approval admin has moved the risky thing first and kept the trivial one.

The five rungs: decide how much authority goes with the work

The single most common delegation failure is not choosing the wrong task. It is handing over a task without saying what the person is allowed to decide. Say “can you take the payroll queries” and you have transferred an activity but no authority, so every genuine decision comes straight back — and both of you conclude the handover failed. Naming the rung out loud, at the point of handover, removes the ambiguity in one sentence.

Two rules make the ladder work. First, name the rung in the same sentence as the task: “you own the payroll queries at rung four — decide and act, and bring anything over £500 of extra fee to me.” Second, move up a rung deliberately rather than by drift, and say when you are doing it. People who are quietly still at rung two after eighteen months conclude, reasonably, that they are not trusted.

Rung three is where most firms should concentrate their effort, because it is where review lives. A threshold rule — jobs under an agreed fee reviewed by a manager, with three named exceptions that always come to the owner, such as a first-year client, a group structure or a disclosure decision — converts a vague worry about quality into a rule anyone can apply. It also makes the size of the problem visible: if the review queue does not shrink after a threshold is set, the constraint was never the owner’s judgement, as the review-queue measure in KPIs every accountancy firm should track tends to reveal.

None of this survives without a written standard. The rung says how much authority; the standard says what “done properly” means. Where a process is genuinely repeatable, that standard belongs in a procedure rather than in a conversation — which is the point of building SOPs in an accountancy practice. Delegating from memory means the standard changes each time you are asked, and inconsistency reads to the team as unfairness.

Worked example: what fourteen hours a week costs and returns

Delegation is usually discussed as though it were free. It is not: it costs real money in the first months, and being honest about that is what stops an owner abandoning it in week six. The figures below continue the illustrative twelve-person firm above. They are arithmetic you can rerun on your own numbers, not benchmark data.

The cost side comes first, because it arrives first. Promoting a senior to take review carries a pay rise — say £6,000 a year in this illustration — and roughly fifteen hours a month of their time moves out of chargeable work into reviewing and supervising. At an illustrative £70 an hour of recovered fee, that is 15 × £70 × 12 = £12,600 a year of chargeable capacity redirected. Add the two of them working through jobs together for the first six weeks, and the true first-year cost is around £18,600 plus a slower autumn.

The return side takes longer to appear and is larger. Fourteen hours a week across roughly 44 working weeks is 616 hours of owner time. Suppose only a third of it — about 205 hours — goes into advisory work at an illustrative £150 an hour rather than into the general relief of a quieter week. That is £30,750 a year, against £18,600 of cost: a net £12,150, with the other two-thirds of the released time going into planning, sales conversations and the firm itself. If none of the released time is redirected deliberately, the honest expectation is that it fills with more of the same work and the return is close to zero. That is the actual risk of delegating without a plan for what the time is for.

Two things about this arithmetic are worth saying plainly. It is negative for the first two or three months in every version of it — the pay rise and the shadowing land immediately, the released hours accumulate slowly. And the largest benefit does not appear in it at all: a firm where review does not queue behind one person delivers work faster, which shows up in cycle time and lock-up rather than in the owner’s diary, as set out in the guide to lock-up.

Why delegated work comes back — and how to stop it

Work returns to the owner through four recognisable routes, and each has a specific fix rather than a general exhortation to hold firm.

• Reverse delegation. Someone brings you a problem and leaves with your solution; the task is now yours again. The fix is a sentence: “what do you think we should do?” — then agree or adjust their answer rather than supplying your own. It costs ninety seconds more the first few times and considerably less thereafter.
• Rescuing under deadline pressure. Two days before a filing date the owner takes the job back “just this once.” Everyone learns that pressure returns work to the top. The fix is to move the internal deadline earlier so there is room to coach rather than rescue, and to treat a missed internal date as a planning finding at the monthly operating review rather than an emergency.
• Clients routing round the new owner of the task. A long-standing client emails the owner directly, the owner answers, and the handover is undone politely. The fix is an explicit introduction — “Priya now runs your payroll and will answer faster than I can” — followed by forwarding rather than answering for at least a quarter.
• Silent standards. The work comes back different from how you would have done it, so you redo it. Unless it is wrong, that is a preference, not a standard. The fix is to write the standard down and then hold the work to that standard only, which also makes it possible for a third person to do the job next year.

Underneath all four is the same test. If you would be genuinely uncomfortable being unreachable for two weeks, the honest reading is that authority has not moved — only activity. Blocking out those two weeks, and letting the firm plan properly for them, is one of the more effective forcing functions available to an owner, and it usually surfaces the remaining dependencies faster than a year of good intentions.

What to do this week

• Log your own hours for the next ten working days, categorised at the end of each day. Do not tidy it up — the untidy version is the useful one.
• Pick the single largest line that does not need you, and hand it over this fortnight with an explicit rung and one written standard.
• Set a review threshold: a fee level below which a manager signs off, plus exactly three named exceptions that always come to you.
• Write down what the released hours are for — advisory work, sales conversations, planning — before they exist. Unallocated time is reabsorbed within a month.
• Put “what came back to me, and why” on your monthly agenda for the next six months. The pattern in that list is the real delegation problem, and it is rarely the one owners assume.

Delegation done properly is slower to start and far more durable than it looks: a fortnight of logging, one handover at a time, a rung named out loud, and a standard written once. Sequencing those handovers so the firm gets easier rather than shakier — and staying with them past the month where it costs more than it saves — is the work described in the Optivo Monthly COO and staged in the Optivo method.