Why does everybody want to "export to CSV"? Is that step still needed today?

Hook

Boss, here is something to notice. Whichever accounting system a company runs — desktop, cloud, expensive, cheap — the first thing the people using it do is almost always the same: "Can I export to CSV?" The reports are right there in the system, so why does everybody still move the numbers into Excel? That is today's subject. And an honest question with it: in this day and age, is that step still needed?

What do they actually do with it?

First, to be clear: they are not being lazy. Across a lot of clients, what they do after exporting comes down to just a few things. One, regroup. The system gives you a grouping by account, but the boss wants to see it "by branch", "by salesman", "by customer source". The system does not have that angle, so they arrange it themselves in Excel. Two, match against data outside. Take the numbers in the books and match them against something that is not in the system — the warehouse's stock count sheet, a list the salesman handed in, a statement the bank gave you. The books are in the system, the other half is outside it. Three, a one-off analysis. The boss suddenly wants to know "over the past three years, how much has this customer bought in total?" Done once, never done again. Four, hand it to somebody else. The auditor wants a copy, the bank wants a copy, the tax agent wants a copy — and every one of them wants a different format. Four things, and only one of them is something the system genuinely cannot do.

Three of them are symptoms

Take them one at a time. "Regrouping" — that is a symptom. Having to rearrange by branch in Excel every month means your books were never recorded by branch in the first place. (The chart of accounts from the fourth lesson, the contacts from the fifty-sixth, and the master data coming in the next episode.) Set it up right and the system gives you the branch report directly, with no export. "Handing it to somebody else" — that is a symptom too. Arranging the auditor's schedule by hand every year means those classifications were not set up to match the format you have to hand over (the forty-eighth lesson). Set it up right once and it comes out directly, year after year. "A one-off analysis" — this one can be exported, but be honest: is it really one-off? If you do the same table "one-off" three months running, it is not one-off any more; it is a regular report you are missing. Only "matching against data outside the system" is where Excel genuinely wins. The warehouse's handwritten stock count, the salesman's list, a sheet somebody else gave you — these things are not in the system, and you need somewhere to put both sides next to each other. For that one, Excel is still the best tool there is.

So what about today's AI?

Now for what everybody wants to ask. Can AI replace this step? Take it apart and the answer differs. What it can already do: the "one-off analysis" category. You used to have to export, build a pivot and write formulas; now you can simply ask, "over the past three years how much did this customer buy, arranged by quarter". It answers. For that part, AI really has done away with Excel. Where it helps, but depends on your setup: "regrouping". AI can only group by the dimensions already recorded in the books. If you never recorded the branch, AI cannot conjure a branch report either. (That line from the thirty-sixth lesson applies just as well here: it is not that AI cannot classify, it is that nobody has told it what your books look like.) What it still cannot replace: "matching against data outside the system". If the data is not in the system, AI cannot see it either. You still have to put both sides next to each other — the only difference is that once they are lined up, AI can help you find which entry the difference is in. So the honest answer is this: AI turns "export then do it by hand" into "export then ask", but it has not made the setup less important. It has made it more important.

A very practical test

Here is a rule you can use straight away. Next time somebody asks you for a CSV, ask one question first: "what are you exporting it to do?" The answer will land in one of three places. One, "I want to see another angle" → that is a setup problem. Tell your accountant to put that dimension into the books. Set it up once and you never have to export again. Two, "I want to match it against something outside" → that is fair enough. Export it, no problem. But put that external data into a fixed format too, and next time will be faster. Three, "I do not know, we have always done it this way" → that is habit. Stop it once and see. Usually nobody notices that the report is gone. And one more thing to say: every export leaves you with one more "offline truth". The moment an Excel file is saved it is detached from the system; the system changes an entry next month and that Excel file will not change with it. Walking into a meeting with an old Excel file is more dangerous than having no report at all.

To close

One line to sum up: most "export to CSV" is not a credit to Excel, it is a debt in the reporting. Only one thing genuinely belongs in Excel: putting what is inside the system next to what is outside it. The other three should all be solved inside the system — and with the setup right, you will find you have not pressed "export" for months. If you are still exporting and rearranging every month, that is not because you are not hardworking enough. It is because nobody has ever set that set of books up properly. Grab a coffee first and talk about your business. For the accounting, come to LTT. I am LTT, helping SME bosses get their accounts straight. Follow us, and see you next time.