What we measure, and what we got wrong

What the dataset actually is

A normalised record of provider list prices — what named companies publicly charge for a named plan — with price points recorded at 2020, 2022, 2024 and 2026 where we have them.

That last figure is the uncomfortable one, so we put it on the page rather than in a footnote: 0 of 161 price records currently carry a primary source URL and a verification date on file. Records without one are labelled as indicative on their own page. Raising this number is the most valuable work we can do on this site, and you can watch it move.

List prices are not household bills

This is the distinction the old survey blurred, and it is the one that matters. Knowing that the median plan we track costs €10.00 a month tells you what companies charge. It tells you nothing about how many subscriptions a household holds, which tier they chose, whether they are on a promotion, or what they pay in total.

Answering those questions needs consumer research. We have not done it, so we do not answer them — and our calculator reports your own total back to you without comparing it to an average we cannot substantiate.

How a figure is produced

1. One plan per provider. Normally the standard individual tier — not the cheapest ad-supported option, not a bundle, not a student rate.
2. Prices are recorded as published. We never convert currencies. A plan priced in dollars is shown in dollars, because a converted figure changes with the exchange rate and could not be reproduced from this dataset.
3. Indices use euro-priced plans only. The basket is 156 of 161 records. Plans in other currencies are excluded rather than converted.
4. We publish medians, not means. A handful of expensive business plans would drag an average upwards and misrepresent the typical price.
5. Price change is computed per plan, then the median is taken — not by comparing two baskets, whose composition changes over time.

Limitations

If you are citing us, cite these too. They are part of the finding.

• These are provider list prices, not household bills. They describe what companies charge, not what anyone actually pays or how many subscriptions people hold.
• The basket is euro-priced plans only. Plans priced in other currencies are excluded rather than converted, because a converted figure would change with the exchange rate and could not be reproduced from this dataset.
• One plan per provider is tracked — typically the standard individual tier. Cheaper ad-supported tiers, student rates and bundles are not in the basket and would lower every figure if they were.
• The basket is not weighted by how many people subscribe to each service, so it is not a consumer price index. A provider with a million subscribers counts the same as one with a thousand.
• Historical points are recorded at 2020, 2022, 2024 and 2026 only. Price changes between those points are invisible to this dataset.

Median list price by category

Categories with at least three euro-priced plans. Reproducible from the dataset.

Automated validation

Every record is checked at build time: currency validity, non-negative prices, whether the annual figure is coherent with the monthly one, whether the five-year figure follows the documented method, whether the historical series is ordered, whether stated inflation matches the stated endpoints, whether slugs are unique, and whether any verification date sits in the future.

Flagged records are not silently corrected. Several of the current flags are VPN and software plans where an advertised “monthly” price is really an effective rate on a multi-year commitment — a real pricing pattern that needs a person to resolve, not a script to average away.

Corrections

If a price here is wrong, tell us and we will fix it and log it. A link to the provider’s own pricing page is the fastest way to get a record corrected: info@innopulse.io. Our full correction log is at /corrections.