A software engineering hiring process works better at 4 rounds than 9.
Cut it to 4. A screen, practical coding, a depth round, a hiring manager conversation.
Run it in 1 day so the panel meets while the memory is fresh.
And require every scorecard submitted before anyone speaks in the debrief.
A typical engineering loop runs 7 to 9 interviews over a month.
It still produces decisions the panel cannot explain a week later.
I defined the hiring process for a team at Zalando that now covers 27 markets. 3 changes did most of the work.
Change 1: why cut the loop to 4 rounds?
Because 7 rounds mostly re-measure the same signal, with tired people doing the measuring.
Four cover it. A 30-minute screen. A 60-minute practical round from your own codebase.
A 60-minute depth or design round. A 45-minute hiring manager conversation.
A fifth round adds noise.
Screen, 30 minutes. Experience, motivation, level. Does this person want this job or any job.
Practical coding, 60 minutes. A problem from the actual codebase, stripped down.
Let them use AI, documentation, whatever they use at work. You are hiring the person who will work that way on Monday.
System or technical depth, 60 minutes. One design problem with clear constraints, scored against criteria written in advance.
Hiring manager and team fit, 45 minutes. How they work with people, how they handle disagreement, what they need from a manager.
Change 2: why run it in 1 day?
Because a loop spread over 3 weeks loses candidates to companies that decided faster.
One day, on site or remote, with breaks, and the panel meets the same afternoon.
Vacancy cost runs around 1,000 euro a day for an open engineering role.
3 extra weeks of process is not free.
The cost is coordination.
You stop losing the people you wanted. Your engineers spend fewer hours in total.
Nobody re-reads notes from 12 days ago.
Change 3: why do scorecards have to come before the debrief?
Because opinions contaminate each other within seconds.
Whoever speaks first sets an anchor. The rest of the panel drifts toward it.
Every interviewer submits score and evidence before anyone speaks.
Including the most senior person in the room.
Then the debrief examines disagreement rather than averaging it away.
This is the change with the highest return and the lowest cost.
What this does to the numbers
Structured assessment has about 0.51 validity against job performance. Unstructured conversation has about 0.20.
Cutting from 9 interviews to 4 saves roughly 5 to 10 engineering hours per candidate.
Across 5 hires that is close to 288 hours back.
You get better decisions and the saving comes along with it.
What about AI in the coding round?
Give them a task that requires it. Watch what they ask, what they accept unchecked.
And how they validate the output.
An engineer who prompts carelessly and ships the result is a real risk. Now observable.
That is a far stronger signal than a binary search written from memory.
Time spent trying to catch candidates using AI is time not spent designing that task.
What do most teams get wrong when they try this?
They change the process without changing the interviewers.
More than 80% of people involved in hiring have had little or no training.
Being a strong engineer does not make someone a strong interviewer.
Shorten the loop without calibrating the panel and you have made a noisy process faster.
Building systems and evaluating people are different skills.
Where to start
Pick 1 change. The scorecard rule is free and you can apply it to the next debrief.
Then look at the loop length, then at interviewer training.
If you want the whole thing done properly, that is what SmartHiring is for.
Process audit at 500 euro, workshop sprint at 2,500.
Common questions
Can 4 rounds really assess a staff-level candidate?
Yes, with the depth round scoped differently. At staff you use the same 4 slots and change what each looks for. Ambiguity handling, choosing what not to build, influence without authority. Adding a fifth round usually means the panel has not agreed what staff means here.
Is a full day of interviews fair to candidates?
Fairer than 3 weeks of evening calls, in my experience of running both. Give real breaks. Tell them the schedule in advance. Pay for the travel. Candidates who struggle with a single day struggle with the fragmented version too. That one wastes more of their time.
How do you get senior people to submit scorecards first?
Make it the rule for everyone including you. Enforce it once in public. The debrief where a director’s score arrives late is the one that sets the norm. Grant exceptions to seniority and you get back the anchoring the rule removed.
Which of the 3 changes should you do first?
Scorecards before the debrief. It is free and needs no scheduling. You can apply it to the next debrief on your calendar. Loop length needs coordination, and interviewer training needs budget. Start with the one that costs nothing and shows an effect in one meeting.
The short version
4 rounds, 1 day, scorecards submitted before anyone talks.
Train the interviewers or the rest is decoration.