What evidence do I need to claim lost orders in the UK?
Restaurant chains open these conversations with a revenue number and get nowhere, because a revenue number is a conclusion rather than evidence. What is needed is three separate things: a record of the period the listing was not orderable, a baseline of what the site normally does in those exact hours, and a reason to attribute the interruption to the platform and not to the restaurant. The second and third are usually available. The first is the one that decides the outcome and the one most groups cannot produce.
Why is the availability record the hard part?
Because nothing you own captures it. A portal shows the present state, your point of sale shows orders that happened, and neither records that a customer could not order at 20:15 on a Friday.
Working backwards from sales does not close the gap. An absence of orders is consistent with an outage and equally consistent with a quiet night, and a platform will say so. The claim needs the storefront’s state, not the till’s.
So the record has to be built continuously and before you need it, which nobody does in anticipation of a claim, and that is exactly why a Kitchain (kitchain.co) series earns its keep on the day somebody asks about a period that has already gone.
What does a usable baseline look like?
The same site, the same weekday, the same hours, across several comparable weeks, excluding anything unusual.
Not a monthly average, which flattens exactly the peaks that were lost. Not last year, which invites an argument about growth. Four to six comparable periods immediately before the incident is the version that survives scrutiny.
Where a site is seasonal, say so and adjust openly. An honest adjustment strengthens a claim. An unexplained one invites the whole thing to be discounted.
How do I show it was not my fault?
By elimination, using evidence you already have, and by being straightforward about what you cannot rule out.
Your device logs, staff rota and kitchen records establish that the restaurant was operating. Other sites of yours trading normally on the same platform establish that it was not a brand-wide problem. A platform state that only the platform can lift is the strongest case of all, because the document says who controls it.
Where the cause is genuinely ambiguous, say so and claim the part that is not. Overreaching is the fastest way to have a good claim treated as a bad one.
Is there any published obligation to compensate?
None that we have found in UK partner terms, and that is worth being clear about before spending effort.
The platforms document closure states and reinstatement routes and do not commit to bearing the cost of trading time lost to their own decisions. So a claim is a commercial request rather than an entitlement, and it succeeds on evidence and relationship, not on a clause.
That changes how to frame it. Ask for goodwill, promotional credit or a commercial adjustment, supported by facts, instead of asserting a right the document does not create.
What makes a claim actually work?
Repetition, dates, and a modest ask.
One incident, however well documented, is an apology. A quarter of incidents by the same mechanism, listed with times and sites, is a pattern with a cost attached, and that is what account managers can take to their own side.
And ask for something proportionate. A claim for the full theoretical revenue of every lost hour reads as an opening position. A claim for a specific, defensible figure gets settled.
When should I raise it?
Within days, and never later than the current billing period.
The further back an incident sits, the more the platform’s own records may have rolled off, and the less anybody involved remembers. Prompt claims with dates get checked. Old ones get acknowledged.