2026 benchmark · gtm engineering survey methodology
Survey methodology
The 2026 benchmark is a self-selected online survey, not a census. The final analysis file has 228 rows and 61 columns. Field work ran from Q4 2025 into Q1 2026, and I published on March 5, 2026. Compensation on this site uses only people who disclosed, and only cells with at least 10 people.
These are 2026 numbers. Help write the 2027 ones.15-minute survey · your comp percentile · 51 free perks from GTM toolsTake the survey →| Measure | Figure | Note |
|---|---|---|
| Rows in the final file | 228 | |
| Columns | 61 | |
| In-house / agency / freelance | 133 / 58 / 30 | |
| Declined salary / equity / bonus | 22 / 18 / 13 | |
| Usable in-house salary rows | 107 | |
| Publish rule for pay | n at least 10 |
Limits
- People who take a GTM engineering survey are not all GTM engineers. Treat this as a benchmark of respondents, not the whole profession.
- Pay has non-response. Twenty-two people declined salary. I cannot correct that bias.
- Some tool rows were grouped (ZoomInfo with Apollo) or noisy (a parser leftover labeled etc, which I dropped).
- Free-text themes are keyword counts. They are not a second census.
- I do not publish respondent-level rows.
2027
The 2027 survey has been in the field since October 2026. It is a new sample. I will not mix those answers into the 2026 tables.
Questions
Why does the homepage say 225+?
That is the rounded public line. The final file I am citing on these pages has 228 rows.
Why hide small pay cells?
The site already promises compensation is only reported for groups of 10 or more. I am holding the new pages to that.
Can I have the raw file?
No. Individual responses stay unpublished. The public dataset is the aggregate tables on this site, plus the report itself through the email gate.










