If you work in DevOps consulting, you make promises measured in days. A pipeline clean-up will take three days. A review of the deployment process will take a week. A hand-over session will fit into an afternoon. Some of those promises land and some do not, and most of us never sit down to see which was which. This post is a checklist for doing exactly that: looking back over a quarter of estimates, setting them beside the real effort, and turning the gap into a better quote next time.
You can use it alone, with a colleague, or with a client who would like to understand why work sometimes takes longer than first said. The tone is relaxed on purpose: make a drink, open a spreadsheet and tick items off.
Before you start: gather what you need
- The original estimate for each piece of work, in the wording you used at the time. A proposal email or a project note is fine.
- The hours or days actually spent. Timesheets, calendar entries and commit history can all help you reconstruct this.
- A list of tasks that were dropped, added or reshaped along the way. These matter more than people expect.
- Half an hour of quiet. This is not a task to squeeze between calls.
If you cannot find the original estimates, that itself is the first finding. Write it down. From today, every number you give should be kept somewhere you can retrieve it.
Step 1: Put estimates and actuals side by side
Make a table with one row per task. Columns: task name, task type, estimated effort, actual effort, ratio. The ratio is actual divided by estimated. Do not judge yet. Just fill it in.
A common surprise at this point is how many rows lack an estimate at all. Small favours and quick questions often never get a number, yet they eat real time. Add them as rows with an estimate of zero and a note, so the overall picture is honest.
Step 2: Sort by task type
Group the rows by kind of work: discovery sessions, pipeline changes, security fixes, documentation, training, hand-over. A single quarterly average hides what you really want to know. Maybe your pipeline work lands close to the guess, while documentation runs at nearly double. That is a very different message from “we are about thirty percent over”.
For each group, write the typical ratio. If you have only two or three rows for a type, mark the figure as tentative and keep collecting.
Step 3: Look for the direction of your error
Ask a simple question of each group: are the misses mostly long, mostly short, or evenly mixed? Consultants tend to underestimate work that depends on other people, such as access requests, approvals and waiting for a team to be free. If most of your ratios are above one, your habit is optimism, and a correction factor will help. If they are mixed, the problem is more likely unclear scope, and you should look at how tasks were defined.
Step 4: Find the three biggest gaps
Circle the three rows with the largest absolute difference between the guess and the actual. For each, write one sentence answering “what did we not see at the start?”. Typical answers include an environment that was less documented than expected, a dependency owned by a different team, or a request that grew once work was underway.
Resist the urge to find a single culprit. The aim is to discover which kinds of hidden steps keep showing up, because those are the ones to ask about during the next scoping call. A guide to how a full engagement unfolds, like this outline of a consulting engagement, moving from initial review to a team that can carry on alone, can help you spot where in the sequence your surprises usually sit.
Step 5: Check how the estimate was made
- Was it given on the spot, or after some thought?
- Was it based on a similar earlier job, or on instinct alone?
- Did it include time for communication with the client team?
- Did someone other than the person doing the work make the estimate?
- Was the estimate ever updated when the scope changed?
Mark the answers in an extra column. Often the quickest estimates, the ones that came out in a call, have the biggest ratios. That finding alone justifies a new rule: no figures in the meeting, a figure by end of day.
Step 6: Calculate a correction factor for each type
Start from each group’s typical ratio and round it sensibly. If documentation has landed at 1.8 and training at 1.1, then your next documentation estimate should be multiplied by about 1.8 before it goes into a proposal, and training by only a little. Keep these factors pinned above the table, in plain view.
Be careful not to double count. If your gut number already includes a cushion, applying the factor on top will inflate the quote. Test it on one or two upcoming jobs and watch the results before you rely on it everywhere.
Step 7: Share the pattern with the client
Clients are usually relieved, not alarmed, when a consultant says, “Over the past quarter, discovery work has taken about a third longer than my first guess, so I have adjusted this proposal.” It signals that the numbers have been checked against reality. It also invites the client to help with the hidden steps, for example by arranging access before the first day.
If you are explaining why you changed your approach to quoting, keep it short and concrete. One table and two sentences are enough. Broader reading on how consulting supports organisations through change, such as this practical piece on how consulting can unlock agility in large organisations, can give the conversation context, but the numbers themselves will carry the message.
Step 8: Decide on one change and one habit
Avoid ending the review with a pile of resolutions. Pick one change to the way you quote, for instance applying the documentation factor, and one habit to the way you record, for instance logging actual hours the same day. Put both in your calendar for the next review in three months.
A note on keeping the habit alive
Reviews fail when they feel like a punishment. Treat the table as a weather record rather than a report card. The point is not that you were wrong; it is that now you know how the weather behaves. Share it with a colleague and ask what they see, because a second reader often notices patterns that the author misses.
Inspiration for this kind of open, regular reflection can come from unexpected places. The website of the Cotocus Foundation presents a nonprofit organisation working with technology, innovation and education, and whose menu has separate corners for programs, research, a community and a blog. A group that writes publicly about its work has a reason to look back on it honestly, and a consulting team can borrow the same instinct for its own records.
The quick version
- Collect original estimates and real effort.
- Tabulate and compute the ratio per row.
- Group by task type and note the typical ratio.
- Name the direction of your error.
- Explain the three biggest gaps in one sentence each.
- Review how each estimate was produced.
- Turn typical ratios into correction factors.
- Tell the client what you changed and why.
Work through the list one time and you will gain more insight into your own planning than a year of vague impressions could provide. Repeat it each quarter and your quotes will quietly become more dependable, which is the kind of reputation that consultants build one honest number at a time.
Best Cardiac Hospitals Near You
Discover top heart hospitals, cardiology centers & cardiac care services by city.
Advanced Heart Care • Trusted Hospitals • Expert Teams
View Best Hospitals