The per-site read. Each site against its own pool of twin stores. This is the number that goes in your pro forma.
EV Charging
Finally — a rigorous way to analyze what retail contributes to EV charging investment.
A fast charge holds a customer on your forecourt for twenty to thirty minutes — about 5x a fuel stop. It is the largest block of captive dwell convenience retail has ever been handed. The industry can neither measure what it is worth nor monetize it properly.
We have a fix for both.
The business case is missing a line
20–30 minA DC fast charge
4–5 minA fuel stop
~5xThe dwell you are now holding
Field observation · charge-lane session timing
Financed today
Electricity sales only
Retail contribution: $0
Payback: 9.1 years
With retail counted
Electricity sales plus radiated retail
Retail contribution: $10,600 per site, per year
Payback: 7.8 years
Every charging business case we have seen is financed on electricity sales alone. The retail a site drives is carried at zero — not because it is zero, but because nobody can put a defensible number on it.
We can. It changes the business case and it shortens the ramp to payback.
Radiated retail: the store sales a charging site pushes into the building around it.
Why nobody has a radiated retail number
The industry keeps trying to follow the driver from the stall to the till. That chase breaks in six places.
- Loyalty opt-in — Members only, and only when they remember to scan. Your best customers, not your marginal ones.
- Cameras — Real money, real privacy exposure, and they see a body, never a basket.
- The charging app — The session data belongs to the network, not to you.
- Walled gardens — The network holds the session, the payment processor holds the transaction, your POS holds the basket. None of the three were built to hand their piece to the others, and none of them want to.
- Free or bundled charging — When the charge is free, or bundled into a lease, a loyalty perk or a hotel stay, there is no transaction at all. Nothing was recorded, so there is nothing to match.
- Asking drivers — What people say they did is not what they did.
A charging session and a store receipt sit in two different companies. Nothing in one points to the other, and no technology is going to change that.
So the industry falls back on sales went up after we installed. That cannot separate a charger from weather, gas prices, the season, or roadworks.
- Charging vendors publish dwell and spend multiples that would transform a site's economics.
- Large operators say publicly that they have looked and found nothing.
- Neither shows how they got there.
Stop chasing the person. Compare the store.
Chasing the person
Needs opt-in, apps or cameras
Members only
Breaks at leased sites
Produces a claim
Comparing the store
Needs nothing you do not already have
Every customer who walks in
Works at leased sites, where you cannot see the sessions
Produces a number with a margin of error
We stop asking which customer did what, and ask what the site did. An install is a dated event at a location.
For each charging store we build a model of what that store would have sold if the plugs had never gone in — built from twin stores that look like it and sell like it, fitted on the years before the install, then run forward. The gap between what the store actually sold and what the model says it would have sold is the retail the chargers drove.
Every other force acting on that store — weather, the local economy, the price of gas, a competitor opening down the road — moves the twin stores too, so it cancels out whether we named it or not.
One site: weekly inside sales and radiated retail
The contribution sits on the modelled baseline. Actual sales move around both.
The green band is the radiated retail the chargers add, sitting on top of what the store would have sold anyway. It is about 0.85% of a week.
The dark line is what the store actually rang. It swings several times further than the band every single week — which is exactly why a before-and-after comparison finds nothing, and why anyone who tells you they eyeballed this number is guessing.
Added up across the year it comes to $10,849 per site. Real, and invisible to inspection.
Simulated · Confidence B · illustrative figures on a de-identified nine-site network. Your own read replaces every figure here in week one.
Three reads, because they fail in different ways
Plug up against plug down. The same store on days the plugs worked against days they did not. The store is its own comparison, so nothing about the store can explain the difference. The sharpest instrument we have.
A bounded demand shock. Where an operator has run a free-charging window or anything like it, we read that too. A change with a known start date and a known end date can be told apart from a season in a way a permanent change cannot.
How we see something this small
- Stop measuring the whole store. Read the categories a charging customer actually buys, in the hours sessions actually happen. The same dollars against a base a third the size.
- Use the broken plugs. Downtime is not a data problem. It is a comparison group.
- Count site-days, not sites. A live network across two years is thousands of store-days, not a handful of locations.
- Get to dollars per charging session. One number that prices every site you build next, rather than a percentage per store that prices none of them.
- Set the noise floor first. We run the same analysis on stores with no chargers, on randomly chosen dates. If it finds something where nothing happened, nothing it finds anywhere else can be trusted.
None of this makes the number bigger. It makes it trustworthy. Making it bigger is the Board's job.
What the number is made of
Once you can measure it, the number stops being a verdict and becomes an equation — and every term in it is a movable lever.
Sessions9 sessions/dayNine is the industry median. Grocery-adjacent sites run to 42. This term has the most headroom.
× Come inside40%Plug-to-door distance, sightline, seating, wayfinding.
× Basket$8Dwell product, loyalty on kWh, local.
= Per site$29 a day$10,600 a year, per site. Simulated.
That is one number. Which denominator you put it over decides whether anyone in the building accepts it.
| Denominator | What it answers | Who accepts it |
|---|---|---|
| Total store sales | "How much did the store move?" — $29 on a $5,000 day is 0.6%. | Nobody. At that size the number reads as a rounding error and the conversation ends. |
| The revenue and P&L the chargers enable | "What is the charger asset actually producing?" — $10,600 a year of store sales, on top of the energy margin, which is the only line the business case counts today. | Site economics and asset owners. It is a second revenue line where there was one. |
| Net capex in the payback case | "Does this change the hurdle?" — against roughly $300K of net capex on a four-stall site after incentives, energy margin alone pays back in about 9.1 years. The store line adds roughly $5,300 a year of gross profit and brings that to 7.8. | Finance, and the capital committee. This is the only version of the number that moves a decision. |
1.3 years off the payback and about 1.8 points of unlevered yield, from a line the pro forma currently records as zero. It is a small number in the first denominator and a decisive one in the third, and it is the same number. A measurement is only useful if it can cross a departmental boundary intact — which is why we hand you all three and tell you which one to walk into the committee with.
Figures here are illustrative, built on the median assumptions above. Yours replace them in week one.
The dwell is barely monetized
Twenty captive minutes on your forecourt. Here is what the industry currently does with them.
- An ad on the charging screen. We have ruled it out on our own field work: drivers look long enough to confirm the session started, then look at their phones.
- A kitchen that closes at nine. The plugs run all night.
- No sign from the stall to the door. The car wash has one. The coffee brand has one. The chargers do not.
- Nowhere to sit. Twenty minutes and no reason to walk in.
The Lever Board is the fix. Every proven move for a charger-equipped store, ranked by what it is worth at your site, tagged with which part of the equation it moves, with its confidence grade and its source — including the moves we tell you to skip, and why.
Rows sort Sessions first, because a site nobody stops at cannot be converted at any rate.
A worked example on a de-identified network. Every figure is illustrative and says where it came from.
Other questions we answer
Can it even be read at my sites?The Detection Floor Check. A one-page answer before anyone spends money.Should we build this site at all?EV charging feasibility. Demand, utilization economics, measured charger-to-store lift.We're underwriting a charging buildout.A capital verdict — the Diligence Read.We operate charge points or sell host sites.Settled reads travel into negotiations as evidence.
Have a different question?Ask it →
What we do not do
Seurat measures what chargers do to the store. Siting, hardware, utility interconnection, and installation economics stay with the partners who do that work well. We are not an EV consultancy and we will tell you when the question you have is theirs and not ours.