Portaltjekby Osmund
Menu

How we review the comparison sites

The method is public, so anyone can repeat the measurement and check our work. Version 1.2. It may be used freely with attribution (CC BY 4.0).

1. What we examined

All Danish websites that show private consumers a comparison, ranking or “cheapest/best” recommendation of electricity retailers. We found them through about 35 different Danish searches (for example “sammenlign elpriser” (“compare electricity prices”), “billigste elselskab 2026” (“cheapest electricity company 2026”), “skift elselskab” (“switch electricity company”), “elselskaber oversigt” (“electricity companies overview”)), through the affiliate networks’ public programme listings (only partly available), the Danish Consumer Council Tænk (Forbrugerrådet Tænk) study of ten comparison sites (7 Dec 2023), a sale listing on Saxis and the ranked sites’ own links to sister sites. The result is 33 sites in the ranking and 4 sites that only collect contact details, plus 5 references that are not comparison sites (members’ test, journalism, app, spot price overview, foreign portal). If you find a site we have missed, write to us.

2. What “whole market” means

The yardstick is elpris.dk, the statutory price portal of the Danish Utility Regulator (Forsyningstilsynet), where all electricity retailers must report all standard products for customers using under 100,000 kWh a year (Executive Order no. 593 (BEK 593/2024)). On 2 November 2026 the portal had 54 companies with at least one priced product for private customers in the Radius grid area (postcode 2200). We use that figure as the denominator in “shown with a price”. A company counts as shown only if the site shows a price for it, not merely its name in a list.

3. How we observed

All sites were read on 29 October 2026 using automated retrieval of their text. 17 sites, including all those that show prices only after JavaScript has run, were also checked in a real browser on 2 November 2026 (elpris.dk, Strømligning, Samlino, Tjekelpris, Elpriser.dk, eltjek24, Elberegner, Strompriser, DinElPortal, Elbob, Sammenlign-nu, Elprisberegner, NicolasKJ, Mikkel Kjerri, Min Strøm, Ekstra Bladet Tilbud and the data feed behind elpris.dk). Dated screenshots have been kept for the sites where the browser was used. The date of the latest check is given on each site’s review.

We counted the priced companies, recorded buttons and links, and read “about us”, “how we make money”, method, privacy and terms-and-conditions pages. Ownership was looked up in the Danish company register (CVR: datacvr.virk.dk and cvrapi.dk) on 2 November 2026 using the CVR number the site itself gives; where a site gives none, we searched by company name, address and email. Foreign entities (Norway, Thailand) have not been checked against a register. The AI test was carried out the same day with one Google search (“hvilket elselskab er billigst i danmark” (“which electricity company is cheapest in Denmark”), Danish setting) and saved as a screenshot.

In all reviews we distinguish between what the site says, what we observed and what we infer; inferences are marked as “presumed”, “possibly” or “not confirmed”. We review sites, not people: people are mentioned only in the role that the site itself or the company register gives.

4. Ten criteria, 24 points

All criteria have a legal basis: Directive (EU) 2019/944 of the European Parliament and of the Council, Art. 14(1)(a)–(h) (requirements for comparison tools), which is implemented in Denmark in the Danish Electricity Supply Act, § 82 b(2), nos. 1–8; and the Danish Marketing Practices Act, § 6(4) (commercial intent must be disclosed clearly), § 6 a (ranking parameters must be disclosed) and Annex 1 no. 11 a (hidden paid placement is always misleading). Private sites are not bound by Art. 14, but it is the standard the legislator has set for what a comparison tool must be able to do. Points are awarded at fixed levels:

1. Ownership and funding disclosed

3 points. Art. 14(1)(b) · § 82 b(2), no. 2

  1. 0No publisher named (no company, CVR or person).
  2. 1Partial: only a person’s name or only a company name; funding unclear.
  3. 2Company and CVR disclosed, and the business model is described.
  4. 3Company, CVR, responsible persons and funding are all clearly disclosed.

2. Advertising labelling

3 points. The Danish Marketing Practices Act, § 6(4) · Annex 1 no. 11 a

  1. 0No labelling.
  2. 1Disclosed only on “about us”, in the footer or in body text at the bottom.
  3. 2Banner at the top of the page OR a label with each paid link/placement.
  4. 3Banner at the top AND a label with each paid link/placement. (Sites without commercial links get 3.) Deduction: if the site calls itself independent/impartial/unbiased while the display or the order is limited to or influenced by partners, 1 point is deducted (but not below 0).

3. Market coverage

4 points. Art. 14(1): “at least one tool must cover the whole market”

  1. 0Under 10% of the 54 companies are shown with a price (0–5).
  2. 110–29% (6–16 companies).
  3. 230–59% (17–32).
  4. 360–89% (33–48).
  5. 490% or more (49–54). Stale prices (over 120 days) halve the score.

4. Honesty about coverage

2 points. Art. 14(1): “a clear statement if the market is not fully covered”

  1. 0Claims “all”/“complete” but shows a fraction.
  2. 1Neither a claim nor a clear statement, or a partial statement.
  3. 2States clearly who is missing, or actually shows the whole market.

5. Equal treatment in ranking

4 points. Art. 14(1)(a) · § 6 a · Annex 1 no. 11 a

  1. 0Only partners are shown and ranked, presented as the market’s “cheapest/best”.
  2. 1A selection (typically partners) is priced; highlighted partner slots at the top.
  3. 2All are shown, but headlines/top lists/awards are reserved for partners, or the default order is an unpublished weighting.
  4. 3All are shown and ranked by price or by published, non-commercial criteria, but only partners/paying companies get a link or a badge.
  5. 4All are ranked by price with the same treatment, including a link, or the option of one, for all.

6. Ranking criteria published

2 points. Art. 14(1)(c) · § 6 a

  1. 0No explanation of the order.
  2. 1A brief statement or a list of criteria without weights.
  3. 2A method page with parameters, weights and price basis.

7. Timeliness and timestamp

2 points. Art. 14(1)(e): “the time of the last update”

  1. 0No date, or prices older than one quarter, or links to companies that no longer sell electricity to private customers.
  2. 1A date exists but is inconsistent, coarse (month only) or attached to stale content.
  3. 2A clear date and signs of ongoing updates within 30 days.

8. Price basis

2 points. Art. 14(1)(c)–(d): “clear and objective criteria, unambiguous language”

  1. 0Unclear or misleading (for example VAT and duties left out without saying so, or fixed and spot mixed).
  2. 1Only the supplier’s part, clearly explained, or the full price without a visible calculation of grid tariff and duties.
  3. 2Full price including grid tariff for the chosen grid area or part of the country (Vest/Øst), with a visible calculation.

9. Option to report errors

1 points. Art. 14(1)(g)

  1. 0No procedure/contact for corrections.
  2. 1An option for corrections or a contact for errors is given.

10. Data minimisation

1 points. Art. 14(1)(h)

  1. 0You must give your name/phone/address to see offers.
  2. 1The comparison can be seen without personal data.

5. Rules that apply across the board

  • Grades: A (“Whole market”) = 20–24 points, B (“A selection, open about it”) = 15–19, C (“Partner-driven”) = 10–14, D (“Critical shortcomings”) = 0–9. The grade names describe the typical picture in each range; the individual points are given on each site’s review.
  • Ties on points are decided by 1) equal treatment in ranking, 2) market coverage, 3) advertising labelling, 4) alphabetical order.
  • “Who can be listed?” is not a criterion but a summary of the observation “does an electricity retailer have to pay to be shown?”: All companies (non-paying companies are also shown), Mostly partners (few without an agreement are shown), Partners only (only paying companies get a price), Unclear (the site does not say).
  • Stale prices: if a site’s prices are older than one quarter (120 days) on the day of observation, the coverage score is halved (rounded down). If a site shows several dates, the newest date next to the price table applies. This affected elspejder.dk (table from 24 March 2026), elprisportalen.dk (June 2025) and ladeportalen.dk (Sep 2025); elradar.dk (6 July 2026 = 119 days) avoided it by one day.
  • Showing Osmund earns a site no points. Whether our company is shown is an observation, not a criterion.
  • No bonus for design, traffic or brand. The two most visited private sites (Strømligning about 590,000 and elberegner.dk about 210,000 visits a month according to Semrush, Sep 2026) are scored on the same criteria as a one-person site.
  • Uncertainty: Each review has a confidence level (high/medium/low). Low typically means that the site is anonymous or that we could not see the whole comparison.

6. What the AI did and what humans did

Searching, reading, counting, CVR lookups, awarding points on the scale above and writing the texts on these pages were carried out by an AI model (Claude, Anthropic) with access to the web and a browser, directed by Osmund’s founders. The scale, the weights and the decision to publish are the work of humans; the person responsible for the content is Niklas Allamand Dib, co-founder and CEO. Before publication, a separate AI run, which had not seen the work being done, checked all owner, model and coverage claims against the sources; it found 14 claims that were unsupported or wrong (including an incorrect description of elpriser.dk’s “cheapest” rule and an incorrect traffic claim about elberegner.dk), all of which were corrected before the first edition. The AI can still be wrong: if it miscounts, or if a site has changed, we correct it and note it in the changelog.

7. Limitations

  • The figures are snapshots from two days. Sites change partners and prices continuously; we repeat the measurement every quarter.
  • Coverage was measured for one grid zone (Radius/2200) and typically one consumption profile (4,000 kWh). A few sites show other companies in western Denmark.
  • Beneficial owners are hidden in CVR for most private limited companies; we state the legal owners and the management.
  • Tænk’s full test from 2023 is behind a paywall; we used only the public parts.
  • The Min Strøm app’s comparison feature could not be checked without the app and has not been reviewed.
  • The AI test is one search on one day in one country; it shows the mechanism, not an average.
  • We have not checked whether the sites’ prices are correct, only whom they show and how they disclose it.

8. Primary sources

Changes and corrections: the changelog.