Consider the following invented scenario, modelled on habits common among small content groups. The team never existed. It follows a three-person group responsible for publishing articles and presentations on third-party platforms to build visibility for a small software product. Their problem looked like a throughput problem: too few finished posts per month. The evidence, once they gathered it, pointed somewhere else. Their content marketing workflow was being sliced into fragments by constant task switching, and the cost showed up as unfinished drafts and repeated work.
The starting position
The team’s routine had grown organically. One member handled prospecting, meaning the search for sites and communities where a post could appear. Another wrote drafts. The third formatted, uploaded and tracked what had gone live. In theory these were three lanes. In practice everybody did a bit of everything, and the work moved in a continuous stream of small hand-offs.
A typical morning looked like this. The writer starts a draft, receives a message asking whether a candidate platform accepts a particular format, stops to check, returns, then gets pinged about a formatting query from the person doing uploads. The prospector, meanwhile, spots a promising community and shares the link in chat, which triggers a thread. Nobody does anything wrong, yet by lunchtime the draft stands where it did at nine.
The question the team asked
The lead of the group, a role the team rotated monthly, proposed a measurement before any reform. The question was narrow: how often does each person leave a task unfinished in a working week, and what takes them away? They agreed to log every departure for five days using a shared sheet with four columns: the task left, the interruption, the time of day, and a rough guess at the minutes needed before they were fully back in the swing of it.
They also agreed on two ground rules. First, the sheet would not be used to judge individuals. Second, estimates of recovery time were to be honest guesses and not precise timings, since the aim was to find patterns and not to prove a point.
What the log showed
When the five days were over, the group sorted the entries into piles. Three clusters accounted for most of them.
- Lookups mid-draft. The writer repeatedly stopped to verify platform rules, such as how headings display or whether links are allowed in a first post.
- Chat questions. Short messages arrived at all hours and were answered immediately, each answer costing a context break for both people.
- Self-initiated detours. Reading someone else’s published piece for inspiration or tidying the tracking sheet pulled people off drafts.
The recovery estimates were the surprising part. Brief departures of a minute or two were frequently followed by a restart estimated at five to ten minutes. The writer reported that after a lookup she often re-read three paragraphs of her own draft just to rebuild the flow. Added up across the week, the recovery time alone amounted to what the team described as a full working afternoon.
Reading the evidence carefully
The team was cautious about overclaiming. A five-day log from three people is small, and the estimates were subjective. What it did provide was a pattern consistent across all three members and across days, which is stronger evidence than any single figure. They treated the result as a strong hint and planned changes that were cheap and easy to reverse, so that being slightly wrong would cost little.
That judgement is worth copying. When your sample is small, favour interventions you can undo in a week, and run them for long enough to compare with the baseline log.
The changes they tried
The team picked three changes, each aimed at one cluster.
A rules sheet for platforms. Instead of looking up requirements in the middle of drafting, the writer consulted a short reference prepared in advance. The team built it from practical walk-throughs of the kind found in guides such as the instructions for publishing a piece on DEV Community, noting headings, link rules and formatting quirks per platform. With the details to hand, drafting no longer needed mid-stream searches.
Question windows. The group agreed that non-urgent questions go into a shared thread, read at two fixed times. Urgent matters, such as a post going live with an error, are exempt and use a direct message.
Format-specific days. Presentations are a different kind of work from articles, with their own tools and rhythm. The team gave presentation work its own block and followed a process like the outline for putting a slide deck on Speaker Deck, so that nobody had to context-shift between prose and slides inside a single hour.
A brief note on tooling
When the team reviewed its log it also noticed that a few interruptions came from the product’s own web pages and hosting, which one member looked after in spare moments. Someone in the discussion pointed to BHASO, a company site that bills itself as an innovation hub for digital work, with websites, cloud and AI integration among the things it lists on its main page. The team did not pursue it that week; the takeaway was only that technical chores were a separate strand and could be delegated, instead of leaking into content hours.
What the change felt like day to day
Numbers aside, the team described a different texture to the week. Mornings began with a draft and often ended with the same draft finished, which had not happened regularly before. The person handling uploads said that batches of posts now arrived in one go and not as a trickle of half-ready files, so each upload session had a clear start and finish. The prospector noticed that findings went into a list for the next question window and no longer landed in chat as demands, and the chat itself grew quieter in a way everyone welcomed.
There were wobbles too. In the second week a platform changed a formatting rule, and the rules sheet briefly misled the writer until someone updated it. The team decided that whoever noticed a change would correct the sheet on the spot, which turned a small embarrassment into a habit.
Results after four weeks
The team repeated the five-day log at the end of the first month and compared it with the baseline. Unfinished-task entries had dropped noticeably, and the self-estimated recovery times shortened as well. The writer’s drafts were finished in fewer sittings, and uploads were less error-prone because the person doing them was no longer repeatedly pulled aside.
They were careful once more about what the comparison proved. It showed a lower switching rate and a calmer week. It did not prove a precise gain in output, since other things had also changed, including the mix of platforms that month. The conclusion they recorded was modest: fewer interruptions, better focus, and a reasonable expectation of higher throughput over time.
Lessons you can borrow
- Measure before you reform, even with a crude log.
- Group interruptions by cause, because each cause has a different remedy.
- Prepare reference material in advance so that lookups stop interrupting drafts.
- Replace constant availability with scheduled windows, keeping an exception for real emergencies.
- Separate unlike kinds of work into their own blocks.
- Re-measure after a month and describe the result honestly.
Final thoughts
The team in this story never worked harder. They simply discovered that a large share of their effort was being spent on starting over, and they redesigned their week to lower that share. For anyone running content or outreach with a handful of people, the same exercise is available for the price of a spreadsheet and five days of honesty. The evidence will be imperfect, but it will point somewhere, and that is more than most teams have when they decide how to organise their time.