# SMM panel price survey — 2 and 13 September 2026 The dataset behind . Fifteen public SMM panel service catalogues, read between 00:16:39Z and 00:16:51Z on 2026-09-02 — a twelve-second window, so every price in here is quoted at the same instant. 44,064 listed services. No account was created, no order was placed, no form was submitted: every page is one a visitor can open without signing in. The same fifteen pages were read again between 20:27:59Z and 20:28:11Z on 2026-09-13, from the same host, with the same script: 44,356 listed services. The two captures are joined on each panel's own service id in `compare.py`. The first capture keeps the file names it was published under; every file from the second carries `-w2`, and its raw pages sit in `pages/w2/`. Two of the fifteen panels — `smmlaunch.com` and `adderpanel.com` — are operated by the company that publishes Novamya. They are in the sample and labelled, because leaving them out would hide the one relationship this data can actually verify. ## Files | file | what it is | |---|---| | `services-all.json.gz` | every one of the 44,064 listed services, as parsed | | `services-basket.csv` | the four basket services per panel, the table the article reports | | `stats.json` | the aggregate figures quoted in the article | | `fetchlog.tsv` | one row per panel: URL requested, URL after redirects, HTTP status, bytes, UTC timestamp | | `method.md` | how panels were chosen, what was excluded, and why | | `pages/*.html.gz` | each catalogue exactly as it was served to us | | `services-all-w2.json.gz`, `services-basket-w2.csv`, `stats-w2.json`, `fetchlog-w2.tsv` | the same four files for the 13 September capture | | `pages/w2/*.html.gz` | each catalogue as served on 13 September | | `changes.json` | every aggregate the article quotes about what changed between the captures | | `price-changes.csv` | one row per listing whose price changed: id, both names, both prices, change in % | | `renamed-listings.csv` | every listing that kept its id and changed its name, so "same listing" can be judged id by id | | `harvest.sh`, `parse_panels.py`, `analyse.py`, `compare.py` | the four scripts, in order. Run them against `pages/` and `pages/w2/` and they reproduce every number offline | ``` python3 parse_panels.py && python3 analyse.py # 2 September python3 parse_panels.py w2 && python3 analyse.py w2 # 13 September python3 compare.py # what changed ``` ## What this is not Listed prices and stated claims on one date. Nothing here is evidence about what any panel delivers, because nothing was ordered. A `$0.01` listing and a `$0.33` listing may not be the same product — that is the finding, not a controlled comparison. `peakerr.com` is in the corpus but excluded from the price-spread table: its `/services` URL redirected to the homepage (see `fetchlog.tsv`), so its rows are advertised headline prices with no service IDs or order limits, and are not comparable with the other fourteen catalogues. ## Licence [Creative Commons Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/) (CC BY 4.0). Reuse it commercially, republish it, build on it; the one condition is attribution. Cite as: > Novamya (2026). *What 15 public SMM panel catalogues listed on 2 and 13 > September 2026.* Dataset. https://novamya.com/data/panel-price-survey/