MERIDIAN CHARGING
Lever Board
Reads updated weekly
Essentially every major automaker has committed to NACS, but committing is not shipping, and the installed plug base is further behind still. Across non-Tesla fast-charging ports, only about 8% have a NACS connector. Parity at the national level is years away.
$11K–$19K per year, midpoint $14.8K, driven by your utilization index of 58.
Transported estimate: connector share is a citation, not a measurementAdapters or dual-cable hardware close this gap at one site well before the industry closes it nationally.
Walked sites show directional signage for the car wash and the coffee brand and none for the chargers. The plugs are invisible from the road at most sites in the archive.
$5K–$12K per year, midpoint $8.4K.
Field-sourced: grade C until the first matched read landsRoad-facing signage and network-map presence. Check that the site is listed and correctly attributed on the major route planners.
Not a lever: a floor. Availability at #11 sits above the twin-site median. The board watches it so a regression shows up as a verdict, not a surprise.
+4.2% charger-to-store conversion at 7 matched sites after stall-side signage and a marked walking path, the cheapest lever on the board and the most consistently measured.
Estimated $2.6K–$4.4K. Measured $3.6K. Inside the range ✓
Validated: our estimate settled on the recordA sign and some paint.
+3.6% charger-to-store conversion at 5 matched sites that gave charging customers a fast lane, the dwell window converts when the line doesn't eat it.
$2.2K–$4.3K per year, midpoint $3.1K, driven by your low conversion index.
Transported estimate: not yet measured hereA signed fast lane, plus a zero-value pass handed out at the stall that rings at checkout as a $0.00 line. The pass costs nothing, so it adds no price effect, and it makes both the use of the lane and its persistence visible in the POS feed. Passes redeemed per site-week then become the dose for a dose-response read across sites. First read at #11 in ~8 weeks.
Sites with tables inside the plug's sightline convert materially better than sites with none. The distinguishing feature is the sightline, not the furniture, which is why the lounge row below is ruled out and this one is not.
$1.6K–$3.4K per year, midpoint $2.4K.
Transported estimateTwo tables in the sightline, this month.
Walked and session-timed at charge lanes in two markets: the driver's clock is set by the charge, and the offer that matches it converts.
Launched Jul 28. The read fills in week by week; verdict lands in ~6 weeks, when 8 weeks of data separate this change from weather and the calendar. Transported midpoint $4.4K.
Read maturing: week 2 of 8Nothing. That's the point: the board watches; you'll get the verdict.
The largest single Capture number on the board, and the only one that cannot be bought with a sign. The offer that matches the charge clock does not exist if the line is closed.
This is a labor decision, not a signage decision. It adds recurring cost and it should wait until the Sessions rows have raised the traffic it would be serving.
Gated: revisit once sessions clear 20/dayRows 1 and 2 landing.
+3.1% attach at 6 matched sites converting charge sessions into store credit; redemption concentrated in the same visit.
$1.9K–$4.0K per year, midpoint $2.8K.
Transported estimateCredit per session, capped. The loyalty row belongs in your app, not on the plug, see the ruled-out row below. This row is also the board's instrument. Loyalty on kWh is the only lever here that ties a session to a transaction without anything from the network, which lets the Capture rows above be read per customer rather than per site. It sees only enrolled members who use the card at the plug, so it is a subsample rather than a census: use it for mechanism and basket mix, and keep the headline lift on the matched-site read.
Built from charge-lane walks and session dwell curves, not a matched-site read yet. Simulated estimate at $1.9K per year: treated as a hypothesis until the first reads land.
Non-alcoholic beer and adult soft drinks are among the few drink categories growing in unit volume while beer overall declines. A charging customer is mid-journey and about to drive, which rules out most of the cold vault's highest-margin shelf.
$0.9K–$2.1K per year, midpoint $1.4K.
Citation: a category trend, not a matched-site readFour facings at eye level in the vault nearest the door. It is a new SKU, so the item master dates the change and the read starts itself.
Session-timed observation is consistent: the driver looks at the unit long enough to confirm the session started and then does not look at it again. Attention moves to the phone for the rest of the dwell.
Coding of screen inventory across 29 charging sites in six countries found half of it given to network self-promotion and under a fifth tied to the shop. That measures what operators put on screens, not what drivers take off them, but it means the inventory is not even being aimed at the basket.
This row used to sit at #5 on this board with a Start now. It moved here when the field read landed. The board is supposed to do that.
Lounges lifted dwell comfort and left store conversion flat at the sites that built them, the spend parks in a room the visit doesn't need.
Against matched sites, inside sales moved within noise while the build-out ran six figures per site.
The build-out you were pricing.
Overnight charging at sites like #11 is pass-through, sessions without store visits. Staffing the counter never covers itself.
At urban nightlife sites, this lever pays. The fingerprint decides: that's the point.
~$48K/yr in overnight labor.
The ruling above rests on a mechanism, not a preference: in app-initiated markets the phone is already in the driver's hand before the session starts. Where regulation puts a card reader on the unit and the driver must interact with the hardware to pay, eyes are on the unit for twenty to forty seconds with nothing else to do. That is a different attention regime, and nothing in the archive tests it. The hypothesis is that screen recall tracks payment modality, not dwell length, which is cheap to test, because it only requires coding sites by how the session is started. Until that read exists this stays an open question rather than a lever, in either direction.