The signals hiding in a UK energy dataset.

The UK energy industry has been quietly opening up its tariff data through public APIs. The obvious use case is the one most people imagine: build a price-comparison tool, find a cheaper tariff. The more interesting use case, which almost nobody is pursuing, is to ask what the whole catalogue says about the suppliers themselves — how they structure it, how their innovation has shifted, and what the data shows that the companies don't talk about.

I recently did exactly that. I pulled eight years of product-level data from the public Kraken GraphQL APIs that EDF and Octopus both run on: more than 15,500 rows of tariff data, every launch, withdrawal, reprice and channel-routing tag across 14 regional electricity zones from 2018 to 2026, with the historical series snapshotted on 13 May 2026 and the live catalogue re-checked in July. This piece focuses on the domestic catalogue. Three findings stood out from the domestic data, none of them visible from the suppliers' websites or mainstream industry analysis, and each one changes how you read the surface of the UK retail market.

Between 2017 and 2021, Octopus invented most of the modern UK domestic tariffs that define the market today. Tracker, its daily wholesale-tracking tariff, launched in 2017. Agile, with dynamic half-hourly pricing, followed in February 2018. Go, for overnight EV charging, arrived in June 2018. Outgoing, for solar export, came in 2019, and Flux, for solar and battery owners, in 2022. Cosy Octopus, for heat pumps, followed shortly after, in late 2022.

Over the same window Octopus also reshaped the market through acquisition: seven legacy customer books absorbed through deals and Ofgem Supplier-of-Last-Resort transfers (ENGIE, Affect, Iresa, Co-op including the Flow book, Tonik, Effortless and Avro), then Bulb in 2022 and Shell Energy in 2023. Alongside that, co-branded deals with MoneySavingExpert and uSwitch effectively turned the comparison sites into Octopus distribution channels.

EDF, over the same period, launched none of those tariff shapes and made no equivalent acquisitions. Its catalogue moved slowly and mostly held a static B2B SME inventory.

Then, in 2024, the pattern flipped. EDF's domestic pricing activity — which had effectively gone to zero between September 2022 and April 2023 — rebuilt steadily over the migration window. By 2025 EDF was pricing and repricing at least as fast as it had in 2021, and at its spring peak faster — though across a narrower range of distinct products than in 2021.

The easy reading is that EDF finally woke up, but that isn't what happened. The real change is that EDF migrated onto the Kraken platform. Kraken Technologies, the back-end Octopus originally built to run its own retail business, was spun out in late 2025 at an $8.65bn valuation with its own CEO. EDF licenses it like any other supplier.

The word Tracker shows how little the names tell you. Octopus Tracker reprices every single day, taking the baseload daily average from yesterday's day-ahead wholesale auction and turning it into one unit rate. Octopus Agile uses the same auction, but passes its half-hourly slots straight through to the meter. Same wholesale market, two ways of handing it to the customer. EDF has sold several products it also calls Tracker since November 2024, but every one of them tracks the Ofgem price cap, which moves four times a year. The same goes for half-hourly dynamic pricing: EDF has experimented with genuine wholesale exposure — FreePhase Dynamic reprices daily against wholesale, and it briefly ran day-ahead half-hourly tariffs for solar and EV owners in 2024 — but these are niche, hardware-gated, and in some cases already withdrawn. Where Octopus makes wholesale exposure a flagship, mass-market proposition, EDF keeps it at the edges of its range.

Octopus didn't only invent the modern tariff catalogue. It also built the biggest of the platforms the rest of the GB market now runs on, then spun it out as a separate company its competitors pay to use. — Two layers of strategic value, from one set of engineering decisions

Kaluza, ENSEK and Gentrack compete for the same licensing layer, but Kraken is the one EDF chose. That is two layers of strategic value from one set of engineering decisions. The product business competes in the UK retail market. The platform business sells to the companies it competes with.

Go to edfenergy.com today and you can buy three tariffs — Simply Fixed, Simply Tracker, and the regulated Standard Variable default (Simply Fixed comes in one- and two-year terms, so four options on screen). EDF actually runs around 30 distinct domestic tariffs — more than 40 if you count each Economy-7 meter variant separately. Octopus is leaner on both counts: two publicly browsable (Octopus 12M Fixed and the regulated Flexible Octopus) against 14 more behind the scenes.

Both suppliers run like a small showroom in front of a much larger warehouse. Most of what they sell reaches customers through specific routes rather than the front door — channel deals at EDF, smart-meter-gated products at Octopus, and for both: retention pages, affinity bundles and migration cohorts.

This isn't deception, it's the ordinary channel architecture of modern utility retail. What it does change is what knowing the market actually means. A consumer comparing what's publicly browsable is looking at five of the forty-odd named tariffs the two leading suppliers actually run. A competitor or analyst doing market research is looking at a curated sample chosen by the supplier's marketing team, and the full catalogue tells a different story about which customer segments each supplier prioritises.

Not every behind-the-scenes tariff is channel-locked. Many, especially at Octopus, are gated to existing customers because the tariff only works on half-hourly smart-meter data: Agile, Cosy, Intelligent Octopus Go, Flux and Tracker all sit in this group. A brand-new customer on octopus.energy effectively can't sign up to the products Octopus is most famous for. Octopus is open about it, describing its standard tariffs on its own site as "the tip of the iceberg." The innovation isn't concealed from the regulator, just gated from the comparison-site economy and held back as a reward for switching in.

The striking part is how little either shows. EDF publicly lists three of its roughly thirty tariffs; Octopus two of its sixteen — both keep the overwhelming majority off the shelf, and EDF hides proportionally even more (about nine in ten, against Octopus's seven in eight). What differs is what they hide, and why. EDF's hidden tariffs are channel-managed variants aimed at brokers, comparison sites, retention pages and affinity partners. Octopus's are gated to existing smart-meter customers because the products only function on half-hourly data. Same shape of catalogue, opposite reasons for the gating.

Since 2024, EDF has launched roughly four times as many fixed-term consumer tariffs as Octopus. People in the industry quote that kind of gap as evidence of EDF's pricing agility, but the comparison is misleading.

EDF builds tariffs around fixed end dates. Each one carries an explicit calendar end date in its code: "Simply Fixed May27v3" runs to May 2027 regardless of when you sign up. As that date nears, EDF has to launch a fresh tariff pointing at May 2028, then May 2029, then quarterly variants, on and on.

Octopus builds tariffs around rolling terms instead. "Octopus 12M Fixed" means 12 months from your sign-up date. The family is conceptually persistent; the clock starts when the customer arrives.

Two things inflate that four-times gap. First, EDF's fixed-end-date architecture spawns more codes per product — one concept like "Simply Fixed" generates a stream of dated variants (~16 codes per concept, against Octopus's ~13). Second, and this is the larger part, EDF simply runs more distinct products. Strip out the end-date churn and count concepts, and EDF still launches two to three times as many fixed tariffs as Octopus. Per concept, both reprice at a similar cadence — so this isn't EDF being more "pricing-agile." It's a bigger, busier fixed-term book meeting Octopus's leaner, rolling-term one.

Architectural decisions look like detail decisions when you make them, and like strategy decisions when someone else has to compete against them five years later.

The fixed-versus-rolling choice sits below the strategy layer most leadership teams ever discuss. It looks like an implementation detail, but it is really an architectural decision taken years ago that still shapes how the company looks to the market: its catalogue size, its reprice cadence, even how regulators read its activity.

Two opposite theories of what energy retail is for.

None of this showed up where you would normally look for it. It was not on the suppliers' websites or in any analyst report, and the companies' own marketing had no reason to volunteer it. The data made it visible, once someone sat down and asked the catalogue different questions.

Set the three findings beside each other and one split runs through all of them. Both suppliers sell the standard predictability products at the front door: a fixed term and the regulated default. The divergence sits behind that door. Octopus has built a separate wing on top, with Tracker repricing daily against the wholesale market and Agile passing the half-hourly market straight through to the meter, reserved for customers who graduate into a smart meter. EDF has dabbled in the same territory but keeps it niche and hardware-gated. Read charitably, those are two different bets about what a household wants from a supplier: shelter from the market, or a path into it once they're ready.

That is the kind of thing a supplier's marketing will never tell you, because it describes a worldview rather than this week's prices. Reading a market like this used to require an incumbent analyst's relationships or a seat inside the supplier; now it mostly needs open data and the patience to ask. As more feeds open, what becomes legible is not just which products each company sells, but what it believes energy retail is for.