First-party research · September 2026
SaaS directory taxonomy benchmark: 340 listings reviewed
This report measures category distribution, automated classification signals, and the editorial controls used to decide which category pages can enter search.
Summary
- 340listing records reviewed
- 19directory categories
- 6high-signal automated suggestions
- 0automatic category changes applied
The classifier flagged 1.8% of listings. Human review rejected each proposed move because the complete product purpose supported the current category. This result shows why a confidence score is a review signal, not an editorial decision.
Largest published categories
These are the five largest primary categories in the published catalog snapshot. Counts exclude secondary-category placement.
| Category | Published listings |
|---|---|
| Productivity | 49 |
| Marketing & Growth | 44 |
| SEO & Content | 30 |
| Developer Tools | 28 |
| Finance & Fintech | 24 |
Method
- Freeze the database snapshot on September 7, 2026.
- Classify each listing from its name, tagline, short description, and full description.
- Flag a change only when title, summary, and combined evidence agree and meet fixed score thresholds.
- Review every flagged record against the product's main customer job.
- Keep category search eligibility separate from listing approval and listing count.
Limits
This is one point-in-time catalog snapshot. Listing text can change, new submissions can alter category counts, and the classifier does not measure product quality, search demand, rankings, clicks, or customer outcomes. The findings do not predict search performance.
Questions and answers
How many listings were included?
The snapshot included 340 listing records across 19 categories.
Did the automated audit change any categories?
No. It flagged six records for review, but a human review kept all six current categories because the full product purpose supported the existing choice.
Does category size make a page indexable?
No. A category needs at least 10 published primary listings and a separate editorial approval for topical quality.