Google Sheets

A finance team asking why stock valuations do not match the warehouse count usually finds a spreadsheet somewhere in the chain. Someone built it to patch a gap between Shopify, the ERP and a 3PL feed, and it has quietly become load-bearing. Cogent² starts by mapping where Google Sheets sits in that chain: which object it feeds, who edits it, and what happens when a formula breaks or a row gets deleted by accident.

That diagnostic tells us whether the spreadsheet is a reporting convenience or a hidden control point for stock, pricing or payouts. The commercial outcome is a scoped build: connect the objects that matter, retire the manual export covering for them, and stop the monthly scramble to reconcile two versions of the same number.

The question that decides whether this goes well is simple: does Google Sheets own anything, or does it only display what another system owns? Treat it as a display layer for stock levels, pricing rules or marketplace feeds and the spreadsheet behaves. Treat it as the place where someone edits a price or overrides a quantity before anything downstream sees it, and ownership gets murky fast, usually days before a promotion or a stock count.

Who owns what, stated plainly: - Stock quantity: the WMS or ERP, never the spreadsheet - Promotional pricing: the ecommerce platform or PIM, with Sheets as a staging view only - Ad hoc reporting: Sheets, pulling read-only data outward

Get this wrong and a buyer edits a formula cell thinking they are updating a report, and the next sync pushes that number into Shopify as live stock. One of the most common support tickets Cogent² sees for this connection starts exactly there.

Point-to-point connections between Google Sheets and three or four other systems age badly: each new marketplace or warehouse addition means another export script, another person who understands it, another thing that breaks quietly when someone leaves. Patchworks replaces that sprawl with a single connection layer built for exactly this kind of object mapping, so adding a new sales channel or a new 3PL does not mean writing another custom script from scratch.

Cost stays predictable because the mapping logic lives in one place rather than scattered across a dozen spreadsheets and scripts. Change is safer too: updating how stock maps from the ERP does not require touching five separate exports. Cogent² operates that layer as expert integrators supported by AI-assisted monitoring, catching quiet failures before they surface as a stock variance at month end, not as a product pitch built on automation alone.

Monitoring a spreadsheet-based feed means watching for absence as much as error. A sheet that has not updated in six hours might be sitting quietly broken, with nobody logging in to notice because the last known values still look plausible. Queue health for this integration means checking whether the expected row count matches what landed, not just whether the job ran. Exceptions worth raising: a batch that completed with zero rows, a mapped field returning blank where a number was expected, a spike in manual edits outside the usual update window.

Catching that before a finance close sees a stock variance saves the two-day scramble of tracing which sheet, which row, which edit caused it. Most brands realise too late that the spreadsheet feed was the quiet failure point, not the ERP or the ecommerce platform everyone was watching.

Training for this integration is short because the risk is behavioural, not technical. The team needs a clear SOP: who is allowed to edit the sheet, what format a new row must follow, and what to do when the sync fails rather than re-running it blind. Runbooks cover the predictable failure modes: blank cells, renamed columns, duplicate rows. Escalation ownership sits with one named person per object, usually whoever owns stock or pricing in the source system, not whoever happens to have edit access to the spreadsheet.

Hypercare after go-live focuses on watching the old manual habit creep back in. Teams revert to spreadsheet patches under pressure, especially in the first peak period after cutover, because the old workaround is familiar and the new flow is not yet trusted. Cogent² holds escalation during that window: triaging failed batches, confirming field mappings against real order volume, and keeping the team from quietly re-opening the CSV export that was supposed to be retired.

Peak readiness means testing the sheet-based feed at volumes the team has not tried before, because a formula that copes with two hundred rows a day can choke silently at two thousand.