土耳其科技薪酬範圍怎樣看：為何中位數及 P25/P75 比平均數更有用

I've spent years in product leadership roles, often on the hiring side of the table, and I've also been the person nervously typing a number into the "expected salary" field. Both sides share the same problem: in Turkey's tech job market, salary information is mostly rumor. A friend's offer, a screenshot in a group chat, a recruiter's "market rate." None of it tells you where you actually stand.

That frustration is why I built Salary Insights. But this post isn't a product pitch. It's about a skill every developer should have, especially early in their career: how to read a salary range correctly, and why the "average salary" you keep hearing about is usually the wrong number to look at.

The problem with averages

Imagine ten backend developers in the same city. Eight of them earn roughly similar amounts. Two of them work remotely for foreign companies and earn several times more.

What's the average? It gets dragged upward by those two people. The resulting number describes almost nobody in the group. The eight "typical" developers look underpaid by comparison, and a junior reading that average walks into a negotiation with an expectation that most companies simply won't meet.

This is the core issue: averages are sensitive to outliers, and salary data is full of outliers. A few very high earners (or a few people entering the wrong number) can shift an average significantly. Salary distributions are rarely symmetrical; they usually have a long tail on the high side.

So when someone says "the average developer salary is X," the honest follow-up question is: average of whom, and how skewed is the data?

Enter the median

The median is the middle value. Line up every salary from lowest to highest and pick the one in the middle. Half the people earn less, half earn more.

The median doesn't care if the top earner makes twice or ten times as much as everyone else. It stays put. That makes it a far better answer to the question most of us actually have: "What does a typical person in this role earn?"

For example, on Salary Insights the current median for Software Engineer is 110,000 TL net per month, based on 5,604 entries. That N matters. A median from five people is an anecdote; a median from several thousand is a signal.

P25 and P75: the range that actually helps you

A single number, even a median, hides a lot. Two roles can share the same median while one has a tight, predictable pay band and the other is all over the place. That's where percentiles come in:

P25 (25th percentile): 25% of people earn less than this. Think of it as the lower end of "normal."
P75 (75th percentile): 75% of people earn less than this. The upper end of "normal."

The space between P25 and P75 is called the interquartile range, and it contains the middle half of everyone in the dataset. It's the most useful salary range you can look at, because it ignores both the extreme low end and the extreme high end.

Here's how I'd read it in practice:

Below P25: You may be underpaid, or your role/level is being classified differently than the market. Worth investigating.
Between P25 and median: Common for people early in a level or newer to a company.
Between median and P75: Solid market positioning, typical for strong performers or people with in-demand skills.
Above P75: You're in the top quarter. Great, but don't use this as your baseline when comparing offers.

Context changes everything: role, experience, city

A range is only meaningful if it's comparing like with like. Mixing a new graduate with a ten-year veteran, or a frontend role with a data role, produces a number that helps nobody.

That's why I'd always slice salary data by at least three things:

Role: Software Engineer, Data Engineer, Product Manager, QA, DevOps, and so on.
Experience: Years in the field shift pay bands dramatically.
City: Where you (or the company) are located still affects compensation.

Salary Insights shows median and P25/P75 by role, experience and city for exactly this reason. When you filter down to your actual situation, the range stops being trivia and becomes a negotiation tool.

Data freshness and sample size

Two more things I check before trusting any salary number:

How recent is it? Markets move. Data from several years ago can mislead you about today. On Salary Insights, out of 53,848 anonymous salary entries, 38,566 come from the last 12 months, so most of what you see reflects the current market.

How many people are behind it? Always look for the N. If a filtered view has only a handful of entries, treat the range as a rough hint, not a fact. Widen your filters (for example, a broader experience band) until the sample is meaningful.

A practical checklist for beginners

Before your next interview or salary conversation:

Find your comparison group. Same role, similar experience, same city if possible.
Look at the median first, not the average.
Note P25 and P75. That's your realistic range.
Check the sample size. Bigger is better.
Decide where you fit honestly. Your skills, interview performance and scope of responsibility determine whether you aim near the median or closer to P75.
Anchor your ask with data, not with a friend's screenshot.

Bonus: let your CV do some of the work

One question I hear constantly is: "Okay, but where does someone with my specific background fall?" Ranges by role and experience get you close, but your CV contains more nuance than any filter.

That's why we added an AI CV salary analysis feature on Salary Insights. It analyzes your CV with AI so you can compare your profile with the median and P25/P75 for your role.

Final thoughts

Salary transparency in Turkey's tech scene is still developing, and that puts individual developers at a disadvantage. The fix isn't more rumors; it's better reading of the data we do have.

Stop asking "what's the average?" Start asking "what's the median, what's the P25/P75 range, and how many people is that based on?" That one shift will make you a sharper negotiator, whether it's your first job or your fifth.

And if you've got a salary to share, add it anonymously. Every entry makes the picture clearer for the next person.

— Kubilay Arkan, founder of Salary Insights & Product Director