Most small projects suffer from one of two planning problems. Either there is no plan at all, just a deadline and a hopeful feeling, or someone builds a detailed schedule in the first week that nobody looks at again by the third.

Both end the same way. Work slips quietly, the client asks for an update, and you find out the real state of things at the worst possible moment.

For a team of three to ten people, you rarely need dependency maps, resource-levelling or critical-path calculations. You need three pieces of information on every task: how much effort it takes, when it starts and when it should be finished. This article walks through how to set those up properly and keep them useful until the project closes.

Step 1: Write tasks as finished outcomes you can actually estimate

Estimates are only as good as the tasks they are attached to. 'Homepage' cannot be estimated. 'Homepage design approved by client' can.

Go through the project and write each task as a result someone could point at and say it is done. Then check the size. A useful rule for small teams is that most tasks should take somewhere between half a day and three days of effort.

  • Smaller than half a day: it is probably a step inside another task. Fold it in, or use a checklist.
  • Larger than three days: it is hiding unknowns. Break it down until each piece is something one person can picture doing.

If you are unsure whether to break a large task into sub-tasks or keep it whole, our piece on when sub-tasks help and when they add noise covers the trade-off in more detail.

Take a generic example: a six-person design agency building a small marketing website over six weeks. The task list might start with discovery workshop held, sitemap agreed, homepage design approved, inner page templates designed, front-end build of templates, content loaded, and launch checklist complete. Seven tasks, each a clear outcome.

Step 2: Estimate effort first, not dates

The most common planning mistake is jumping straight to dates. Someone asks when the homepage design will be done, and the answer is 'end of next week' without anyone working out how much work it involves.

Separate the two questions. First ask: if you could work on nothing else, how many hours would this take? That is the estimate. Dates come later, once you know who is doing it and what else is on their plate.

A few habits make estimates more reliable:

  • Ask the person doing the work. Your estimate as the project manager is a guess. Theirs is informed by having done similar work before.
  • Think in ranges, then record one number. If the designer says 10 to 16 hours, ask what would make it 16. If those risks are likely, record 16.
  • Keep the buffer visible. Rather than padding every task quietly, keep estimates realistic and add contingency at the project level, such as a few unassigned days before the deadline. Padding hidden inside tasks tends to get used up whether it is needed or not.
  • Include review and rework. A design task that needs client feedback is not finished when the first draft is. Estimate the round of changes too.

In Tilvin, each task carries an estimate alongside its assignees, priority and status, so the number lives with the work rather than in a separate spreadsheet.

Step 3: Turn estimates into start and end dates

Now convert effort into a calendar. This is where most of the realism comes in, because a 16-hour task is almost never finished in two working days.

For each task, work through these questions:

  1. Who is doing it, and how much of their week goes to this project? If your designer spends about half their time on this client, 16 hours of work takes roughly four working days, not two.
  2. What has to happen first? The front-end build cannot start until templates are designed. You do not need a formal dependency diagram; just make sure the start date comes after the thing it depends on.
  3. Where are the waiting periods? Client approval often takes days. Give it its own time on the calendar, or set the end date of the design task to include it.
  4. Are there fixed dates? Holidays, a client's board meeting, a launch tied to a campaign. Plan backwards from these.

Set the start and end date on each task once you have answered these. Then step back and look at the whole picture. The project Gantt chart in Tilvin's reports lays tasks out against the calendar, which makes it easy to spot two large tasks landing on the same person in the same week, or a gap where nothing is scheduled at all.

The team workload report is worth a glance at this stage too. If one person is clearly overloaded across all their projects, the dates you just set will not hold, however carefully you estimated.

Step 4: Keep the plan honest with a 15-minute weekly check

A plan stops being useful the moment it stops matching reality. The fix is not more planning up front; it is a short, regular check.

Once a week, ideally on the same day, go through the active and upcoming tasks and look at three things:

  • Time logged against the estimate. If the team logs time on tasks, you can compare hours spent with the original estimate. A task at 14 of 16 hours that is only half done is an early warning, not a surprise next week.
  • Progress versus dates. Ask the assignee to update the task's progress. A task due Friday that is at 20 percent on Wednesday needs a conversation now.
  • Dates that need moving. When something slips, move its end date and check what depends on it. Leaving old dates in place makes the whole plan look fine until it suddenly is not.

Keep notes on what went over and why. After two or three projects, you will see patterns: client reviews always take longer than planned, or build tasks are consistently underestimated by a third. That is the information that makes your next estimates better.

The project health report pulls much of this together. If you want a routine for reading it, see Read Your Project Health Report Before the Client Does.

What you can safely skip

Part of minimum planning is knowing what to leave out. For most small projects, you can skip:

  • Hour-by-hour schedules. Day-level dates are precise enough.
  • Estimating every tiny task. If a task is a 20-minute admin item, a date is enough.
  • Detailed plans for later phases. Plan the next three or four weeks in detail and keep later work as rough, larger tasks. Break them down when they get closer and you know more.
  • Re-planning from scratch. Adjust dates and estimates as you go instead of rebuilding the schedule every time something changes.

Your next steps

You can apply this to a project you are running today, without starting over:

  1. Rewrite your current task list as outcomes, and split anything bigger than about three days.
  2. Ask each assignee for an effort estimate on their open tasks, in hours.
  3. Set start and end dates that reflect each person's real availability and any waiting time for approvals.
  4. Look at the plan on a timeline and fix overlaps or overloaded weeks.
  5. Book a 15-minute weekly check to compare logged time, progress and dates, and adjust.

That is the whole method. It takes an hour or two at the start and a quarter of an hour each week, and it gives you something most small projects lack: an early, honest view of whether you will hit the date.