Customer acquisition cost in a market where switching is rare and slow

There are two ways to put half a million households on a supply book, and they price two orders of magnitude apart. You can buy them one at a time off a comparison table for a fee in the tens of pounds, or you can buy the book: Octopus Energy has absorbed customers in blocks of hundreds of thousands, from Co-op Energy in 2019 to Avro Energy under Supplier of Last Resort in 2021, Bulb in 2022 and Shell Energy Retail in 2023. Both are acquisition. Only one of them lands in the marketing budget, and it is usually the smaller number.

Blend a £45 comparison-site fee, a £90 paid search cost and a £130 door-to-door commission into a single CACCACCustomer Acquisition Cost (CAC) is the total sales and marketing spend divided by the number of new customers gained in a period. It measures how efficiently you grow.View full definition → of £70 and you have a figure that is arithmetically correct and useless. It tells you nothing about which channel to cut, which to scale, or why the door knockers keep hitting quota while margin quietly erodes. Everything below rests on the slow, trigger-driven switching decision the foundations lesson describes; the costing consequence is that spend and acquisition almost never land in the same reporting period.

Why blended CAC lies to you

CAC (customer acquisition cost) is total acquisition spend divided by customers acquired in a period. In energy the numerator and denominator come from different quarters. A comparison-site click today can convert eight months later, after three more visits and two paid search sessions. Last-click attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → hands the entire cost to whoever touched the customer last, usually the comparator at the moment of switching, so upstream brand and search spend reads as inert.

Each route hides its cost somewhere different.

Paid comparison. uSwitch and its peers charge suppliers per completed switch, normally with a clawback if the customer cancels in the cooling-off window or the transfer fails. uSwitch earns its revenue from exactly these supplier commissions, so the fee is visible and gets over-blamed. The invisible part is table position: the ranking is by price, so the discount you cut to sit in the top few results is an acquisition cost wearing a pricing label. Shave £15 a year off a dual-fuel tariff to hold that position, keep the customer two years, and you have spent £30 that nobody booked to marketing, on top of the switch fee.

Paid search and display. Platforms report on last-click or 7-to-30-day windows, far shorter than the real consideration cycle, so the channel is chronically under-credited and then chronically under-funded.

Door-to-door. Agents are paid per acquisition or on commission, and the cost per signed contract looks cheap until you add early cancellation. D2D-acquired customers in several European markets show materially higher first-year attrition than digitally self-selected ones, per industry commentary from analysts such as Cornwall Insight, a UK energy market analysis firm. If half churn inside 12 months, the true cost per retained customer doubles.

Book acquisition. Buying a supply book, or taking one on through a regulator-run process, moves acquisition out of marketing and into corporate development. Under the UK's Supplier of Last Resort mechanism the acquiring supplier takes the customers with no purchase price and recovers honoured credit balances and transfer costs through a levy on all consumers' bills; in a negotiated deal, someone pays a price per relationship. TXU Energy's parent Vistra has grown its US retail base the same way, buying whole books, at implied prices per customer in the hundreds of dollars rather than the tens.

The failure mode is arithmetic. Drop a 580,000-customer SoLR intake into the same denominator as a quarter of paid mediapaid mediaVisitors arriving via paid ads or sponsored placements, where you pay a platform to display your message rather than earning visits organically.View full definition → and blended CAC collapses towards zero, making every paid channel look efficient the month before you cut it. Keep the book in its own ledger.

Building a true CAC: the fix

1. Attribute over the consideration window, not the click window.

Use multi-touch attributionmulti-touch attributionA method that distributes conversion credit across all marketing touchpoints in the customer journey, rather than crediting only the first or last interaction.View full definition → with a lookback matched to your actual sales cycle, 18 to 24 months, not the platform default of 30 days. Most teams skip this because it means stitching CRMCRMCustomer Relationship Management: software and strategy to manage and analyse customer interactions throughout their lifecycle.View full definition → (customer relationship management) records to ad platform data over a horizon no e-commerce playbook contemplates.

2. Report CAC per retained customer at 12 months, by channel.

Not CAC per signed contract. This is the single highest-value fix.

3. Load the cost properly.

Media and agency fees, comparison-site commissions, D2D commissions, sales ops overhead, creative production, and the tariff discount you conceded to buy table position.

4. Keep book acquisitions on a separate line.

Price paid (or levy-recovered cost) plus migration cost: system transfer, the contact-centre spike, complaint handling, divided by customers still supplied a year after the transfer date.

Worked example

A UK supplier's quarter, with illustrative retention rates for demonstration rather than published figures:

Raw CAC makes door-to-door and paid search look identical at £100. Divide by the retained base and D2D nearly doubles to £182, the worst of the four, while the comparison site, usually written off as commodity volume, delivers the cheapest retained switcher.

CAC_retained = Total loaded channel spend / (Customers acquired × 12-month retention rate)

The book row is the one to argue about. It is cheapest per head and you cannot buy one on demand: it arrives when another supplier fails or sells, on tariffs you did not price, sometimes with inherited debt and a meter estate you have not read. Treating £19 as a planning assumption for next year is how a growth target gets built on an event nobody controls.

What "good" looks like: benchmarks

Energy retail CAC is not reported the way SaaS (software-as-a-service) CAC is, and suppliers treat these numbers as commercially sensitive. Directional estimates only:

• UK comparison-site acquisition commonly sits in the £40 to £80 range depending on tariff type and season. Direct switching through brand search runs lower per unit but needs sustained brand spend behind it.
• In ERCOT (Electric Reliability Council of Texas), where TXU Energy competes, residential switching is far more active than in Britain and door-to-door and telemarketing persist; participants estimate loaded per-acquisition costs above $150 to $200 once commission structures are counted.
• Ofgem (Office of Gas and Electricity Markets, the UK regulator) has reported annual domestic switching in the high single digits to low double digits as a share of customers, low against telecoms or insurance. See Ofgem's retail market data for current figures.

The edge case worth planning for: between late 2021 and 2023, wholesale prices sat above the UK price cap, no supplier wanted new customers, and the comparison tables had no live fixed deals to show. The paid-comparison channel did not get expensive, it stopped existing. A CAC model with no scenario for a channel's denominator going to zero will simply return an error where a budget decision should be.

The retention link you can't ignore

CAC only means something next to the lifetime valuelifetime valueLifetime Value: the total revenue (or profit) a customer generates throughout their entire relationship with your business.View full definition → model this module builds elsewhere, and the handover is where teams lose money. Each acquired customer needs a cohort date, a channel tag, a loaded cost and a flag for whether they switched in or arrived by book transfer. Then the arbitration: book the tariff discount once. If the discount sits in the LTV margin line and not in CAC, the channel looks better than it is. If it sits in both, you have counted it twice and you will kill a channel that works. With residential gross margingross marginGross margin is the share of revenue left after subtracting the direct cost of producing goods or services, expressed as a percentage of revenue.View full definition → per customer running in the tens of pounds a year, a £182 retained CAC needs several years of tenure before it clears, which is a long bet in a market where you get one shot per household every few years.

🎬 [VIDEO: "Customer Acquisition Cost Explained" - youtube.com/results?search_query=customer+acquisition+cost+explained - a primer on CAC fundamentals and common measurement pitfalls, useful as a base before applying the energy-sector adjustments above]

Key Takeaways

• Blended CAC hides which channel pays; break it out by route before any budget decision, and never let a book transfer sit in the same denominator as paid media.
• Report CAC per retained customer at 12 months. Door-to-door usually shows the lowest raw CAC and the highest true CAC; comparison sites often invert.
• The discount you concede to rank on a comparison table is acquisition cost, not pricing. Book it once, in CAC or in margin, and never in both.
• Book acquisition prices customers in the tens of pounds per head on migration cost, but it is opportunistic, arrives with someone else's tariff mix and debt, and cannot be scheduled into a growth plan.
• Treat every figure here as an estimate: suppliers do not publish channel-level CAC, and Ofgem publishes switching rates, not acquisition costs.