A shared Google Sheet with frozen headers, a colour scale on CPA and one tab per client is a real reporting system. It took work to build, the client knows where to find it, and it answers questions. The spreadsheet vs dashboard debate usually gets framed as whether the sheet looks professional enough, which is the wrong test. Most of these sheets look fine. The useful question is how much of the sheet lives in one person’s head, and what happens to the client’s view of your agency on the month that person is unavailable.
I have seen agencies run eleven clients off one workbook and I have seen agencies switch to a reporting dashboard tool at three clients and never regret it. The difference was rarely the number of clients. It was whether the monthly fix list was shrinking or growing.
A spreadsheet is a dashboard until the person who built it is out sick
Every mature client sheet has undocumented behaviour. The conversions column on the retail account is pulled from a different export than the one on the lead gen account, because the platform renamed the column in spring and nobody updated the header. The CPM on tab four is calculated from a different date range than the one above it, for a reason that made sense in March. There is a hidden helper tab that nothing visibly references but which breaks three charts if you sort it.
None of this is bad work. It is the normal sediment of a file edited once a week for two years by someone who knows what they meant. The problem appears the first time somebody else has to produce Tuesday’s numbers. They open the workbook, find a cell that reads 4.9% where last month read 2.1%, and have no way to tell whether that is a performance story or a formula that lost its range when a row got inserted.
That is the actual dependency. Not the spreadsheet, the person. And it is invisible until it is urgent, because the author never experiences the ambiguity.
Conditional formatting looks live; the data under it is from last Tuesday
Colour reads as authority. A cell shaded green tells a client that something checked that number and approved of it. In a manual sheet, the only thing the colour confirms is that the value sitting in the cell crosses a threshold, whenever that value was last pasted in.
This gets expensive in a specific way. A client opens the sheet on the 14th, sees green on pacing, and plans around it. Spend has in fact slipped behind since the 9th, which the account manager knows and intends to mention on the call. The sheet quietly contradicted the conversation before it happened, and now the account manager is explaining why the report said one thing and they are saying another. That exchange costs more trust than the pacing miss did.
Dashboards are not automatically immune. A dashboard that nobody refreshes is a stale dashboard with better typography. The difference is that in a sheet, staleness has no shape. There is no timestamp a client would naturally look at, no obvious mark between the part that was updated this morning and the part that was last touched in August. Google Sheets reporting makes freshness a matter of discipline, and discipline is the thing that fails first in a busy month.
One tab per client stops scaling long before client number ten
The tab-per-client pattern works beautifully at three clients, holds at six, and starts eating hours somewhere after that. What breaks is not the file size. It is that any change you want to make to the layout now has to be made in every tab, by hand, consistently.
Say you decide the first screen should lead with cost per acquisition rather than impressions, which is usually the right call. In a sheet, that is a layout change repeated eight times, with eight chances to drag the wrong range. Two months later the client who was never migrated asks why their report looks different from the one their colleague at another brand gets, and you realise two tabs are now two different report designs.
Campaign restructures do worse damage. A media buyer splits one DV360 line item into four by audience, the export gains rows and loses a naming convention, and formulas that referenced fixed ranges either break loudly or, worse, keep returning a number that is now wrong. Pivot tables survive this better than direct cell references, which is why the sheets that last tend to be the ones built on pivots by someone who thinks like an analyst. Most agency sheets are not built that way, because they started as a quick favour for one client.
The structural answer is to separate the layout from the data, which is exactly what report templates do: you decide the layout once, and every client report inherits it instead of imitating it.
The client can always find the formula bar
This is the part agencies underestimate. If you share a sheet, the client sees the machinery. They can click a cell and read the formula. They can see the tab you forgot to hide with last quarter’s working numbers. They can see the comment where someone wrote “check with Sam before sending.” On an editable share, they can also sort a column and break the sheet without knowing they did it, then ask you why the chart looks strange.
Access is the other half. A tab is not a permission boundary. One mis-set share setting, one forwarded link, and client A has a route to client B’s spend. Sheets do have permissions, and careful agencies use separate files per client to contain this, which then multiplies the maintenance problem from the previous section. There is no arrangement that gives you both one place to maintain and hard separation between clients. That trade is built into the tool, and it is the reason the questions a client’s IT team asks about report links and access keys are hard to answer well from a spreadsheet.
What a spreadsheet still does better than any dashboard tool
Honesty matters here, because the switch gets oversold.
A spreadsheet is unbeatable for ad hoc work. When a client asks a question nobody anticipated, you can answer it in four minutes with a formula, and no dashboard will match that. It is also the right place for reconciliation against the media plan, for the messy intermediate arithmetic between what the platform reported and what you will invoice, and for any analysis you are still figuring out the shape of. Nobody should build a dashboard for a question they have asked once.
Spreadsheets are also free of onboarding. Every account manager already knows how to use one, and so does every client. That is worth real money and it is the honest reason most agencies stay on spreadsheet client reports longer than they planned.
So the sensible end state is not a dashboard replacing the sheet. It is the sheet moving behind the curtain, used for working out and for one-off questions, while the client sees something that was designed to be read.
Spreadsheet vs dashboard: where the line actually is
Three signals, and the first one carries most of the weight.
The fix list is growing month over month. Track it for two cycles. If the same five repairs recur, those are a workflow problem you can document. If new repairs appear each month, the sheet is accumulating faster than you are maintaining it, and that curve does not flatten on its own.
Someone other than the author has to produce a report. Holiday cover, a hire, a resignation. The moment the sheet has to be legible to a second person, the undocumented behaviour becomes a cost you pay in reconstruction time.
A client asks for something the sheet cannot give without a new tab. Their own branding on the report. A view that only they can open. A version they can send upward without your working notes attached. These requests signal the client has started treating your report as a deliverable rather than a shared workspace, which is usually the real moment a client reporting dashboard starts paying for itself. The portal versus attachment comparison covers what changes once the client stops receiving files.
Client count alone is a poor trigger. A freelancer with four clients who reports weekly and rebuilds a creative breakdown every time has a stronger case for switching than an agency with nine clients on a quarterly cadence.
If you do switch, keep the upload motion. The export you already pull from DV360 or Meta Ads stays the input, because asking a team to change where their data comes from is how a rollout stalls. In Screebl the Excel or CSV export goes up once per report, the columns become the dimensions and metrics you group by, and the rates (CTR, CPC, CPM, VTR) come off the totals rather than averaging row-level percentages, which is one of the quieter arithmetic errors that sheets invite. The client gets a branded link or a portal where they only see their own reports, and you can unpublish it at any time. You keep the sheet for the questions you have not had yet.
If you are weighing the options more broadly, the buyer’s guide to client reporting software goes through what to check before committing, and the plans are listed here once you have decided the maintenance has outgrown the simplicity.
