
Two engineers, same company, same slow release process. The first rolls up her sleeves and starts fixing things straight away. The second spends a few days documenting how releases currently behave before touching anything. A quarter later, both say their work paid off. Only one of them can show it. Here the two habits get weighed against each other on effort, confidence, career value and risk, ending with a rule of thumb for when the slower start earns its keep.
Approach A: change first, judge later
Acting immediately has real attractions. Momentum is high, stakeholders see activity and small fixes land quickly. For a trivial, low-risk tweak, this is perfectly reasonable.
The weakness appears when you try to judge the result. Memory of the old behaviour is fuzzy, and it is shaped by the last few painful incidents. Someone says the process feels quicker; someone else disagrees; both are right about their own experience. With no recorded starting figures, the argument has nowhere to go, and it usually ends with whichever person speaks most confidently winning.
Approach B: record the starting point, then change
The measure-first route front-loads a modest amount of work. You trace a few recent changes through the release process, note the elapsed time of each step, list where people wait and write down the complaints you hear. It feels slow because nothing visibly improves while you do it.
The payoff comes later. Every subsequent decision rests on evidence, discussions shift from opinion to observation, and you can retire ideas that had no effect without anyone losing face.
Side by side: effort
Approach A costs almost nothing up front. Approach B costs a few days of observation, which is a small fraction of most improvement projects. Where B pulls ahead is in avoiding wasted effort: teams that skip measurement often spend weeks optimising a step that was never the bottleneck.
Side by side: confidence in the result
On this criterion the distance between them is largest. A shows a story; B shows a comparison. When a manager asks whether the investment was worthwhile, A can only offer enthusiasm. B can put two dated pages on the table and say what moved, what stayed put and what got worse.
Side by side: career value
For people building a career, the difference compounds. Engineers who can demonstrate measured improvements stand out in interviews and performance reviews, because they speak in outcomes rather than activities. The route described in the progression from junior to senior DevOps engineer rewards exactly that kind of evidence, and so does the move covered in the guide to moving over from sysadmin or helpdesk duties into DevOps. A portfolio line that reads “release lead time fell after I changed X, measured before and after” beats a list of tools every time.
Side by side: team morale
People notice when improvement work never seems to land. Under A, colleagues may wonder whether the effort achieved anything, and sceptics grow louder with each unmeasured change. Under B, the team sees concrete movement, even small movement, and that visible progress builds goodwill for the next round of changes. A recorded gain, however modest, is easier to celebrate than a vague impression.
Side by side: risk
Approach A hides regressions. If a change speeds up deployment but quietly raises support tickets, nobody notices, because nobody recorded ticket volume. B catches this because the baseline includes the neighbours of the thing you are changing. The risk of B is different: analysis paralysis. Some people measure forever. Cap the exercise at a fixed number of days and move on.
Where certifications fit
The same comparison applies to learning. People who sit certifications often ask which credential to pursue, and the overview of the DevOps certifications that can speed up a career is a fair starting map. But a map does not tell you your own starting position. If you want to know whether a course changed how you work, note beforehand how you currently handle a typical task, what takes you longest and which parts you avoid. That is a smaller, more personal version of Approach B, and it turns a certificate from a trophy into an experiment.
A parallel outside engineering
Careful planners in other fields follow the measure-first logic naturally. A family arranging eye care abroad, for instance, might use BestEyeHospitals, a planning directory for international patients, to organise questions before any enquiry goes out. The site frames its own pages as general information, with a disclaimer that readers should confirm details with providers and not postpone professional care on the strength of it. Write your starting facts down first, then compare whatever you are told against them.
A short decision guide
- Pick A when the change is tiny, reversible and cheap, and nobody will ask you to prove it worked.
- Pick B when the change touches many people, costs real money or time, or will need defending later.
- Use a hybrid when pressure is high: run the quick fix but spend one afternoon capturing the current numbers first.
The practical bottom line
Most improvement efforts deserve the second approach, and the extra cost is small. Capture the route a change takes, the time each stage consumes, the spend and the complaints. Set a deadline for finishing the exercise. Then change one thing and repeat the same observations.
The two engineers from the opening ended up in very different places. One had a pleasant sense of progress. The other had a page of before-and-after evidence she could attach to a promotion request. Both worked hard. Only one of them left a record.