# FDE numbers count job titles. A new survey wants to measure the work

> FDE Hub's State of FDE 2026, open until 9 October, asks practitioners directly and locked its analysis segments before launch.

Original: https://fdetimes.net/en/news/state-of-fde-2026-survey-closes/

Every FDE figure in circulation measures job titles, not work. FDE Hub makes that charge on the opening page of its State of FDE 2026 survey, which accepts responses until the end of 9 October.

For a developer thinking about moving into forward deployed engineering, this matters for more than method. Published FDE numbers shape what you expect from the job: how much time you spend with customers, who you report to and how far your responsibility runs. If the data measures the wrong thing, your expectations will be wrong too.

## Why 1,500 responses are still not enough

The largest benchmark so far is Perspective AI's survey, published in May 2026, which covered 1,500 FDEs at 154 companies with data collected between February and April. Two findings stood out: FDEs spend a median of 47% of their week with customers, and 78% of respondents expect their team to double in size or more in 2027.

In an analysis published in August, Tech Scoop warned that this should be read as an industry survey, not a census of the FDE workforce. The headline salary figures reflect the top of the market rather than the whole profession, and the 78% figure records what respondents expect, not hiring anyone has actually observed. The difference matters: expected growth is not an open role.

The problem goes beyond sampling. Paraform, a recruiting firm, notes that most FDEs never appear on job boards. If recruiters cannot find them by job title, statistics that count by title are unlikely to capture the whole profession.

A simple test: does your current title include "FDE"? If not, how different is your daily work from an FDE's?

## Ask the people doing the work, and fix the questions first

FDE Hub's fix sounds simple: ask the people doing the work. The survey focuses on what a job title cannot reveal: how time divides between discovery and building, who is responsible for the system after handover, reporting line, and salary band.

**Key point:** A survey can only be trusted when its authors cannot change the questions after seeing the answers.

The more notable detail is the method. Milos Mandic, who runs the survey, has committed to analysis segments that were fixed before the survey opened and will be published with the report. The analyst gives up the option to wait for the results and then slice the data in whatever way produces the best headline.

According to Paraform, the profession exists because the last mile of enterprise AI is code written inside a customer's messy environment. That work does not fit neatly under one title, and it should not be measured by title either. If you do this work, 9 October is your last chance to add one more practitioner from Vietnam, or anywhere else, to the data on the profession.

**Try this week:**

- If you do FDE-style work under a different title, spend 10 minutes on the survey before the 9 October deadline.
- Rewrite one line of your CV around the work itself: what share of your time is spent with customers, and what you own after handover.
- The next time you read any FDE statistic, check three things: who was asked, how many were asked, and whether the segments were fixed in advance.

## Sources

- [The State of FDE 2026 Survey Is Open](https://www.fdehub.org/p/the-state-of-fde-2026-survey-is-open)

- [State of Forward Deployed Engineering 2026: Survey of 1,500 FDEs](https://getperspective.ai/blog/state-of-forward-deployed-engineering-2026-survey-report-1500-fdes/markdown)

- [Why Forward Deployed Engineers Are Becoming the Delivery Layer for Enterprise AI](https://techscoop.substack.com/p/why-forward-deployed-engineers-are)

- [How to hire a forward deployed engineer in 2026](https://www.paraform.com/insights/forward-deployed-engineer-demand-quadrupled)
