Linnworks reporting pain: the case for asking instead of exporting

Nobody dislikes reporting. What people dislike is being the human join between a report that nearly answers the question and the question they were actually asked.

What the export cycle really costs

Count it properly. Find the right report. Set the date range. Run it, wait, download. Open it in a spreadsheet. Filter. Realise one column you need is in a different report. Run that one too. Join them by SKU. Read the answer. Type the answer into an email.

Twenty minutes for a question that takes four seconds to ask and one sentence to answer. Do that four times a week and it is a working day a month, spent on the clerical wrapper around the thinking rather than the thinking.

And there is a second cost nobody books: the questions that never get asked because they are not worth twenty minutes. Those are frequently the useful ones.

Where reports are still the right tool

This is not an argument against reporting. Scheduled reports win decisively when the same figures need to arrive at the same time every week for the same people, when the output must be identical month on month so it can be compared, and when the thing is going to an auditor who wants a document rather than a conversation.

Also: anything where indexing delay matters. Processed-order reporting in Linnworks indexes on a delay, so a figure that must reconcile to the penny against a period end wants the report, not a live read. Related: why stock levels look wrong.

Where asking wins

Ad-hoc, one-off, exploratory, and anything with a follow-up in it. “What did we refund last month?” is a report. “What did we refund last month, and is it concentrated in one supplier’s lines?” is a conversation, and the second half is the half that changes a decision.

It also wins on the questions that are embarrassing to ask a colleague for the third time. People will ask software something twice as often as they will ask a person.

What it looks like in practice

Someone types a question into the assistant they already use. The connector calls the right Linnworks endpoints, the figures come back live, and the answer appears with the detail named — the actual orders, the actual SKUs — so it can be checked rather than believed.

Then the follow-up. Then the one after that. The whole exchange is shorter than running one report.

The habit change

The tooling is the easy half. The hard half is that people have learned not to ask, because asking used to mean waiting. It takes a few weeks of answers arriving in seconds before the ops team starts asking the questions they had quietly given up on.

The way to accelerate it: pick the three questions your team currently asks by email and answer them live in front of everybody. That does more than any training session.

If the export cycle is the bottleneck, the fix is asking Linnworks questions instead of exporting them — a connector in your own Cloudflare account, reading live. See also: managing price lists by CSV without descending into chaos.

Questions, or want a tool we don't have yet? Email hello@grafto.co.uk — a real person replies.