Baseline the hours you spend on admin for two weeks, tag them by category, then re-measure at week eight. Count both hours handed over and hours removed entirely by automation.
The honest answer to whether the hire worked is a number. Without a baseline, you get a vague sense that the week feels lighter, which is not enough to justify the cost to anyone else.
Before the assistant starts, log time in rough thirty minute blocks against a handful of categories: diary, inbox, travel, reporting, chasing, data entry. Precision is not the point, pattern is.
Hours alone can mislead. Also watch response times, follow-ups that slipped, and how often meetings are prepared for properly. Those move first and they matter.
Week two is too early, the assistant is still learning the business. By week eight the diary and inbox are stable and the first automations are running, so the numbers reflect the steady state.
You cannot measure hours returned if you never measured the hours spent. A two week baseline is enough, and it does not require a time tracking tool. Tag calendar blocks and keep a short daily note of the admin that happened outside the diary.
Categorise as you go. Aggregate hours tell you the size of the problem, but the breakdown tells you what to hand over first.
Hours handed over and hours removed are different measures and only one of them compounds. Handing scheduling to an assistant returns your time but the work still exists. Automating the reporting that sat behind it removes the work permanently.
Track both. Early on, most of the gain comes from handover. By month three, the removal number should be growing faster.
Week eight is the right checkpoint. Early enough that a bad fit can be corrected, late enough that the first automation has landed and the novelty effect has worn off.
Repeat the same two week exercise using the same categories. If the total has not moved by at least five hours a week, the problem is usually scope rather than the person, and the fix is a sharper brief rather than more time.
This sounds like an afterthought and is in fact the part most often mishandled. Hours returned without a plan get absorbed into more meetings and a slightly fuller inbox, and three months later nobody can point to what the support actually changed.
Decide in advance where the time goes. For most founders and executives the honest answer is one of three things: customer conversations, hiring, or the strategic work that has been postponed for a quarter. Naming it before the hours appear makes the reallocation deliberate rather than accidental.
It is also worth protecting the recovered time structurally. If eight hours a week come back and none of them are blocked in the diary, the diary will fill them with whatever asks loudest. Blocking two half days for the named priority is a crude mechanism that works better than intention.
Then review it alongside the hours measurement. At week eight, alongside the question of how many hours were returned, ask what those hours were spent on and whether the answer is one you would defend. That is the number that eventually justifies the arrangement to anyone else in the business.
How many hours a week do clients typically get back? It varies with the role and how much recurring admin exists, which is why we recommend a two week baseline before the assistant starts so the result is measured rather than estimated.
When should we review the engagement? At week eight. By then the diary and inbox are stable and early automations are live, so the numbers reflect normal running.
Does automation reduce the hours we need from an assistant? Often it shifts them. As routine admin is automated, the same hours move to higher value work rather than disappearing.
More insights | Alvara Partners