EV Charger CMS Features
A charging platform is usually sold as a charger dashboard, which is the smallest thing it does. Five different parties depend on it, they want entirely different things from it, and the one that decides whether the business works is the driver: a driver who does not come back is a site that underperforms, and a site that underperforms is a host who does not renew.
What the platform owes each of the five parties it serves, the tickets it should raise without being asked, the voice agent that answers with real data behind it, and how money moves through it.
Review your requirementsFor An operator or manufacturer writing platform requirements, or replacing a platform that only ever served one of the five parties properly.
The five parties a platform serves
Most platforms are built for one of these and bolt the others on afterwards, which is why the others always feel like an afterthought.
Swipe to compare
| Who | What they need from it | What they do if it is missing |
|---|---|---|
| Drivers | Find, start, pay, see history, reserve | Charge somewhere else |
| Site hosts | What their chargers earned, invoiced and reconciled | Decline to renew |
| Installers | The commissioning checklist and the units they are responsible for | Leave no record behind |
| The chargers | Monitoring, fault handling, maintenance history, knowledge base | Fail repeatedly for the same reason |
| Operators and administrators | Everything above, scoped by who is allowed to see it | Run the network on spreadsheets |
The table is the argument of this page. A platform that handles chargers beautifully and hosts poorly loses sites at renewal, and no amount of charger telemetry fixes that.
Drivers
The party a network cannot afford to lose, and the one most platforms treat as a mobile app problem rather than as a platform problem.
- 1
Find and see
Where the chargers are, which are free now, what they cost before arriving rather than after leaving.
- 2
Start and stop
From the app, from a conversation, from an RFID card, from a call. The platform should not care which.
- 3
Reserve
A bay held for an arrival time, with the rules about what happens when they are late decided in advance rather than argued about afterwards.
- 4
Pay
The gateway, the receipt, the invoice, the refund when a session failed. A failed session that still charges the driver costs more than the session was worth.
- 5
Their own history
Sessions, spend, and where they charge. This is also what makes a battery report or a personalised offer possible at all.
- 6
Be recognised
A driver on their eighth session is a different customer from one on their first, and the platform is the only thing that knows.
Site hosts
The party that signs the agreement and decides whether to sign it again. Hosts churn quietly: they do not complain, they just decline at renewal.
- 1
Their chargers, and only theirs
Scoped so a host sees their own sites in full and no other host's at all.
- 2
What was earned
Sessions, energy and revenue on their units, in a form they can check rather than take on trust.
- 3
Reconciliation
What the platform recorded against what actually settled, so the monthly conversation is a review and not an investigation.
- 4
Invoices and billing
Generated, delivered and correct, including the revenue share arrangement and the tax treatment for their market.
- 5
Utilisation they can act on
Not a chart. Which hours are busy, which bays are not earning, and whether that is demand or a fault.
A host who can answer their own questions without emailing you is a host who renews. It is the cheapest retention there is.
Installers
The party platforms do not build for, which is why so many sites have no usable record of how they were commissioned.
- 1
The installation process, as software
Step by step through wiring, cabling, earthing, accessories and the photographs, rather than a paper form that ends up in a drawer.
- 2
A checklist that produces data
Each step recorded as it is completed, so the commissioning record is queryable later. This is what makes a fault diagnosable two years on.
- 3
The units they are responsible for
Scoped to their own jobs, across sites and across customers.
- 4
Maintenance work
The job, what was found, what was done, the parts used, and what should be scheduled next.
The chargers themselves
- 1
Station management
Live state, session state, and the site conditions around the unit where station hardware is fitted.
- 2
Faults, with enough to act on
What happened, on which unit, what ran before it, and what it did the last three times. Everything the platform cannot tell you remotely becomes a site visit, and site visits are what a network costs to run.
- 3
Asset records
Serial, model, firmware, location, warranty, and the fact that units get moved and swapped.
- 4
Maintenance history as a pattern
One unit failing is an incident. The same unit failing four times is a decision about replacement.
- 5
A knowledge base for mixed fleets
Fault codes mean different things on different makes, and a technician needs the procedure for the unit actually in front of them.
- 6
An OCPP gateway
So chargers from different manufacturers, on different OCPP versions, connect to one platform without each one becoming a special case.
Operators, and who is allowed to see what
The administrator's view is every other view at once, which makes scoping the feature rather than an afterthought.
- 1
Roles that mean something
A host, an installer, a support agent and a network administrator need genuinely different access, not the same screen with buttons hidden.
- 2
Multi-tenancy
Several operators, or several hosts under one operator, sharing infrastructure without sharing data.
- 3
An audit trail
Who changed a tariff, who restarted a charger, who issued a refund. This matters most on the day an auditor asks.
- 4
Support with the context attached
An agent answering a driver should not be opening three systems to see one session.
Tickets the platform raises by itself
The single most valuable thing a platform does, and the one most often missing. A ticket should exist because the network noticed, not because a driver was annoyed enough to call.
- 1
A charger that has stopped talking
No communication for a defined period is a ticket, whether or not a driver has tried to use the unit. Nothing else will tell you.
- 2
A site that has stopped earning
A location with no sessions for a period it normally charges in is an operational ticket in its own right. It may be a fault that does not report, it may be access to the site, it may be the location itself. All three are worth knowing and none of them appears as an error.
- 3
Grouping duplicate events
Ten events from one outage are one ticket with ten pieces of evidence, not ten tickets.
- 4
Categorised on creation
Sorted into the kind of problem it is, and routed on that, so the operations team is working a queue rather than reading a log.
The second one is the case operators most often have no answer to. A charger reporting healthy at a site no driver has used in three weeks is not a healthy site, and only the platform is in a position to notice.
A voice agent with the platform behind it
A support line automated without access to the platform can be polite and nothing else. The value is entirely in what it can see.
- 1
It answers from live data
The session, the charger state, the payment record, the driver's own history. Most support calls are questions the platform can already answer.
- 2
It can act, once the caller is authenticated
A remote start or stop, scoped to that caller's own session.
- 3
It covers the unstaffed hours
Which is when a large share of charging happens, and when a support line is otherwise a voicemail box.
Money, and connecting to other networks
- 1
Tariffs that express what you actually charge
Time bands, user groups, session and idle fees, minimum charges, and the correct tax treatment for the market. Every platform does per-kWh pricing. Far fewer do a real tariff.
- 2
Offers and discounts
Targeted at a driver, a site, a group or a time of day. This is the instrument for bringing a driver back, and a platform that cannot express an offer cannot run a retention programme.
- 3
Transactions and settlement
What each session was worth, what was authorised, what settled, and what did not, held so the three can be compared.
- 4
Payment gateways
More than one, because markets differ and a gateway that is down is a site that cannot take money.
- 5
Roaming over OCPI
So drivers from other networks can charge on yours and yours can charge elsewhere, with the settlement between operators handled rather than reconciled by hand.
- 6
Revenue share
The host's cut, calculated from the same session records the host can see, which is what keeps that conversation short.
The number that decides whether a network compounds
Everything above resolves into one question: does a driver who charged once charge again.
A site earns from repeat traffic, not from first visits. A driver whose first session failed, or who could not pay, or who was never given a reason to return, is a driver the site paid to acquire and did not keep. Enough of those and the site underperforms, the host looks at the revenue share and declines to renew, and the network loses a location it already paid to build.
Which is why churn belongs in a feature list. The platform is where a first session becomes a second one: it is what notices the failed session, what follows it up, what carries the offer, and what tells you whether any of it worked. A platform that cannot do that is a charger dashboard with a payment screen attached.
Want this applied to your own site?
Review your requirementsTechnically reviewed by Deepu Joy, Director of Products and Delivery. Last reviewed 2026-09-10.
Frequently asked questions
What do most platforms get wrong?
They serve one party properly and the rest as an afterthought. Usually the chargers are handled well and the host is handled with a monthly spreadsheet, which is how sites are lost at renewal without a complaint ever being registered.
Why should the platform raise its own tickets?
Because otherwise a ticket only exists when a driver was annoyed enough to call, which means the faults you know about are a biased sample of the faults you have. A charger that stops communicating, or a site that stops earning, will never generate a call at all.
What is the most under-specified feature?
Tariff modelling. Every platform supports per-kWh pricing. Far fewer can express a real tariff with time bands, user groups, idle fees and correct tax treatment at once, and fewer still can carry a targeted offer.
Do we need OCPI roaming?
If drivers from other networks should be able to use your chargers, or yours should work elsewhere, then yes, and the settlement between operators is the part worth checking rather than the connection itself.
How do we test these in a demo?
Ask for your own tariff to be configured, not a sample one. Ask to see what a host sees. Ask to see a fault from last month and what it says. Then ask how a ticket gets raised when no driver has called.
Writing platform requirements, or replacing one that only serves the chargers?
Send us your tariff, your fault history and what your hosts currently receive each month. Those three say more about what you need than a feature list does.
Review your requirements