Virtual Try-On Statistics 2026

01 / Original research

How many recorded clothes try-ons reached a finished result?

Of 120 clothes try-on generation records created from 1–30 September 2026 UTC, 101 ended with a completed or seen status and 19 ended with a failed status. These are backend job records, not unique customers or purchases.

The question comes from shoppers wanting a usable preview, but this count is not a customer success rate. The records do not distinguish internal tests from customer jobs, and a finished image is not proof of visual accuracy, fit or satisfaction.

We also checked whether we could report generation time. None of the 101 finished records in this sample had a usable completion timestamp, so we are not publishing a speed statistic.

02 / Independent research

What published experiments found

• In a June 2026 study with 66 participants, people randomly assigned to an AR fashion try-on group reported an average purchase-decision score of 4.04, versus 3.14 in the non-AR group. The study used a simulated shopping environment and a small sample; it did not test TryAnyOutfit or prove real purchases. The linked journal summary does not state when the participant sessions took place.

  Shalih, Nawangi and Sabrian, 2026 · original study

• A 2019 apparel study combined a 415-person online experiment with 49 qualitative respondents. In its tested interfaces, AR virtual try-on was rated less enjoyable than conventional mobile-shopping presentations and less convenient and useful than model photos with similar physical features. This is a finding about those interfaces, not every try-on product. The linked publisher abstract does not report the fieldwork dates.

  Plotkina and Saurel, 2019 · original study

Both outside studies evaluated augmented-reality experiences. TryAnyOutfit generates a new image from uploaded photos. Their results give context, not a performance claim for our product.

03 / Method

How we counted our records

We queried generation records created from 1 September 2026 00:00 UTC to, but not including, 1 October 2026 00:00 UTC. Of 159 generation records across modes, 120 had the clothes try-on mode label tryon-max. We counted each record once by its last stored status: 96 seen, 5 completed, and 19 failed. Other modes were excluded.

We did not count a user as a distinct person, infer a purchase, inspect photos, or calculate a satisfaction or fit score. We could not remove internal test activity because these records do not carry a reliable test flag. Status may change after the extraction date. For this reason, do not use these counts as a forecast of a shopper’s chance of success.

The calculation is reproduced by our read-only research script. Its private, anonymized status extract contains no user IDs, photo links or personal information. The extract is retained internally so we can audit the published count.

04 / Maintenance

Review history

1 October 2026: First publication. Added September 2026 operational counts, methods and two original studies. We will check the source links and dated figures quarterly and record substantive changes here. The page URL stays the same when the title year changes.