Disputes in ERCOT's taxonomy, not in a shared inbox.
MarkeTrak work is unglamorous, deadline-bound and expensive to get wrong. Conduit models all 10 case types and 47 subtypes with the fields each one actually requires, and runs the inadvertent gain loop end to end.
Off your desk
You stop tracking disputes in a shared inbox and hoping somebody remembers the one that ages out this week.
Without this
What this looks like today.
MarkeTrak issues tend to live in an inbox and a spreadsheet. Somebody owns them in the sense that they are the person who remembers. It works until the volume rises, or that person is on leave, or an issue ages past a deadline nobody was counting.
The subtypes are the other trap. Each one wants particular evidence and particular fields, and the way most teams learn which is by having a submission rejected. That knowledge ends up in one person's head rather than in the process.
Inadvertent gains are the case that costs real money. A customer you gained by mistake is a customer you are serving, billing and about to be charged for — and unwinding it means a dispute, a regain transaction, and someone tracking whether it actually completed.
How it works
MarkeTrak, in detail.
Issue resolution that matches ERCOT's own taxonomy, so a case you open is a case they recognise.
The full taxonomy, with the fields each subtype needs
All 10 case types and 47 subtypes, each modelled with the fields its own workflow requires and the wire codes ERCOT uses. A case you open is shaped like a case ERCOT recognises, so the requirements are enforced when you raise it rather than discovered when it comes back.
Inadvertent gain and loss as a loop, not a ticket
The dispute opens the case; the regain 814 goes out to the market from inside that case; the market's response comes back and lands on it. Gaining and losing sides are both modelled. The loop closes itself, and what is left for a person is the judgement, not the tracking.
Clocks that start themselves
Every case carries its SLA and its aging, visible before it becomes urgent rather than after. Assignment routes work to a person, and unworked cases surface as unworked rather than as absent.
The backlog you are migrating in
Open cases import in bulk with the same mapping, validation and dry run as the customer book. Moving to a new system does not mean keeping the old one open for the tail.
At a glance
- All 10 case types and 47 subtypes, each with the fields its workflow actually requires
- Inadvertent gain and loss handled as a loop: dispute opens the case, the regain 814 goes out, the result comes back
- SLA clocks, assignment and aging so nothing sits unworked
- Bulk import for the backlog you are migrating in
Connected
What this leans on.
Nothing here is a separate product with an integration between it and the rest. These are the capabilities this one shares a record and a ledger with.
Questions
The ones we actually get.
Which MarkeTrak case types does Conduit support?
All 10 case types and the 47 subtypes beneath them, each with the fields and wire codes its own workflow requires — not a generic ticket with a type dropdown on it.
What is an inadvertent gain, and how does Conduit handle it?
An inadvertent gain is a customer who ended up on your service without a valid enrollment — usually a switch submitted against the wrong ESI ID. Conduit models both the gaining and losing side: the dispute opens a case, the regain 814 is submitted from inside it, and the market response lands back on the case. The customer, the case and the transaction chain stay attached to each other throughout.
Can we migrate cases that are already open?
Yes. Open cases import in bulk with column mapping, validation and a dry run, so the backlog moves with the book rather than trailing behind it in the old system.
Is MarkeTrak part of the core platform?
No — it is an opt-in module. Retailers small enough that disputes are genuinely rare do not pay for it, and switching it on later makes it appear in the product and on the invoice at the same time.
The rest of the platform
One customer record, one ledger, one audit trail.
See it against your own book.
Bring a month of real transactions and a TDSP invoice. MarkeTrak is easier to judge against data you already argue about than against a demo tenant.