The First Baseline Isn't Financial: Where Your New Team's Hours Actually Go

Last Updated: 2026-09-11

Before you redesign a single workflow in the business you have just bought, measure where the team's hours actually go. Ask people where their time goes and you will get a confident, largely fictional answer, because interrupted and fragmented work is precisely the work nobody remembers accurately. A time audit, logged in real time and classified against your investment thesis, produces the baseline every later first-100-days decision depends on: it tells you whether the problem you inherited is capacity, process, or a handful of activities that should be delegated or dropped.

The Estimate Gap: Memory Is Not a Baseline

Why start here rather than with the P&L? Because the accounts tell you where the money went, and the owner of a small business can usually reconstruct that in an afternoon. What the accounts cannot tell you is which activities actually consumed the working week, and on that question the people inside the business are systematically wrong about their own days.

The evidence for this is not soft. According to field research by Gloria Mark and her colleagues at UC Irvine, published in 2005, workers took an average of 25 minutes and 26 seconds to return to an interrupted task, and worked through an average of 2.26 other activities before they got back to it. Think about what that does to anyone's recollection of their week. If a task you remember as "the morning's real work" was actually assembled out of fragments separated by two unrelated detours, your memory of how you spent your time is not approximate, it is structurally misleading.

The practical consequence for a new owner is that you cannot interview your way to a time baseline. People will tell you they spend most of the week on their core job, because the core job is what they remember and what they would rather discuss. The gap between that story and the logged reality is exactly where your first operating decision lives, and it is invisible until somebody measures it.

How to Run the Audit

What does a time audit actually involve in an acquired business, done well enough to act on? Two things decide whether the result is usable: catching time in the act, and classifying it against something. Run it as a defined exercise with an agreed window, announce it plainly, and state the purpose before you start: this is a measurement, not a performance review. If the team believes the output is a redundancy list, the data will be defended rather than recorded, and you will have spent the audit window manufacturing a fiction with timestamps.

How long the window should be is a choice, not a fixed rule. Two working weeks is a sensible default when you want individual logs, long enough to smooth out a launch week or one unusual customer crisis. But we have seen audits run successfully as a one-day workshop with the key team members in a room together, reconstructing and classifying the work end to end, and for many small businesses that format is the better starting point.

In fact, we recommend the workshop format over instructing a different individual to run the audit on everyone's behalf. The discussion, the ideation, and the process of hearing other team members' experience and pain points first-hand are genuinely valuable to a business owner in control of process efficiency, particularly hearing directly from the people involved in the end-to-end process rather than reading about it second-hand. There is a second-order benefit too: from what we see, people involved early in this process and in the ideation are more likely to adopt AI agents later, or become champions for them, because they helped name the problem the agent is being aimed at.

Log in real time, not from memory

Have each person record what they are working on as they work, in blocks of 15 to 30 minutes, for the full audit window. A shared sheet, a calendar with honest entries, whatever the team will actually keep up. The tool matters far less than the timing: end-of-day recall produces the estimate gap described above, end-of-task logging does not. One senior person's log will always be less reliable than the whole team's, so capture everyone, including the people whose work you have not yet formed a view on. In the workshop format, the same discipline applies to reconstruction: work from the calendar and the actual recurring commitments in the room, not from memory alone.

Classify against the thesis, not against a template

Once logged, sort every block into three buckets: work that directly advances what you bought the business for, work that is necessary but creates no differentiation, and work that should not be happening at all. The highest-priority bucket is the first one: the activities vital to moving value through the chain, the work that customers actually pay for and that everything else exists to support. Those are the lines you protect, staff properly, and eventually automate with care.

The sorting conversation is worth more than the sheet it produces. From experience, the act of discussing, analysing and documenting how value flows end to end through a business is a valuable exercise in itself: it highlights blind spots that are invisible in daily operations, exposes the real impact of bottlenecks nobody has named, and listening to which items people argue over and which buckets things land in shows you exactly how much of the operation still runs on the owner's personal control. That is diagnostic work no template does for you. The classification has to be yours, because it encodes your thesis. A compliance process the acquirer considers essential is a very different object from one the previous owner ran purely out of habit, and no generic template can tell them apart. This step is where the audit stops being a measurement and becomes a decision list, and it is the step most likely to start an argument. Let it. The argument is the diagnosis.

Triage: Three Categories and What Each One Means

What do you do with the results? Harvard Business Review research by Julian Birkinshaw and Jordan Cohen gives the benchmark and the method. Across 45 knowledge workers in 39 companies and 8 industries, they found people spent on average 41% of their time on activities that offered little personal satisfaction and could competently be handled by someone else. When 15 executives applied a structured triage to their own logs, they cut desk work by an average of six hours a week and meeting time by two hours a week, roughly a working day recovered, filled instead with work only they could do.

Their triage sorts activities into three types, and it maps directly onto an inherited business:

The order matters. Quick kills first, because categorisation lets you make immediate impact in the business: the dead wood is eating away at your margin every week it survives, and stopping it costs nothing and frees time today. As a business owner you want to know where your money is going and what return each activity returns for it, and this is the step that makes that visible before it is lost sight of entirely. Delegation second, because it moves work without changing the process underneath it. Redesign last, because changing a process you have only just measured is how new owners break something they did not know was load-bearing.

One thing the triage deliberately does not do is point at technology yet. Only once every activity is categorised do you get a good idea of the potential areas where AI could add value, and the answer is never the dead wood. It is the high-value items, the key constraints sitting on the work that moves value through the chain, where a change to how the bottleneck operates is worth more than any amount of tidying elsewhere. Running an automation decision before the triage means aiming it at whatever was loudest in week one rather than at the constraint that actually limits the business.

Who Owns Each Activity Afterwards

Every line that survives the triage needs a name attached to it, and for the first hundred days the discipline is simpler than it sounds: one owner per activity, stated in writing, including the ones that stay exactly as they are. The audit gives you the list. What turns it into an operating baseline is a record that says, for each recurring activity, who performs it, who is accountable for its quality, and what happens when that person is absent. In a founder-run business that question usually has an unspoken answer: the founder. Post-acquisition, that person is you, and you cannot personally be the exception path for every recurring activity.

Synergised's own engagements start from the same place: before any recommendation about tooling or automation, the first deliverable is a documented picture of how work actually flows today, because a redesign aimed at an imagined process is worse than no redesign at all. The time audit is the smallest version of that discipline that a new owner can run unaided in the first month.

When the Audit Misleads

A measured baseline can still be a wrong one, and it is worth naming the three ways this goes off before you trust the sheet. First, an unrepresentative window: if the log lands across a payroll run, a stock take or one large customer escalation, the picture will overweight firefighting and understate the steady state, so check what the audit period actually contained before classifying. Second, gaming: when people sense the output ranks them, logged blocks drift towards the work that flatters, which is why the exercise is announced as a measurement with no performance attachment and repeated at the same cadence each cycle. Third, classification drift: the same activity can land in different buckets depending on who sorts it, so the triage needs one classifier, or a written definition per bucket that everyone applies. None of these argue against the audit; they argue against treating one window as settled truth. Where the result surprises the team, a second log in a different week settles whether it was real.

The Audit Sheet Is the Evidence

Keep the raw log and the classified baseline. It is easy to see the exercise as disposable, something you run once and act on, but its second life is more valuable than its first. When you later cut a workflow's handling time, delegate a monthly process, or drop a report nobody read, the audit sheet is the dated before-measure that proves what changed, by how much, and on whose authority. It becomes the first entry in a benefits record for the acquisition, and the reference point for every measurement that follows: you cannot show an improvement against a baseline you never took.

Sources

• Birkinshaw, J. and Cohen, J., "Make Time for the Work That Matters", Harvard Business Review, September 2013. https://hbr.org/2013/09/make-time-for-the-work-that-matters
• Cohen, J. and Birkinshaw, J., "Make Your Knowledge Workers More Productive", HBR blog, September 2013. https://hbr.org/2013/09/make-your-knowledge-workers-more-productive
• Mark, G., Gonzalez, V. M. and Harris, J., "No Task Left Behind? Examining the Nature of Fragmented Work", Proceedings of CHI 2005. https://doi.org/10.1145/1054972.1055017