Why the numbers can differ

Your analysis is a snapshot of one moment. Your dashboards are live. Over time the two drift apart, and that is by design rather than a fault.

The short version

Each weekly and monthly analysis is written once, from the data that had reached GrowInsight at the time. It is then kept exactly as written. Your dashboards, meanwhile, keep updating as new information arrives from your practice management system.

So if you open an analysis from three months ago and compare it against the same period on your Financial dashboard today, some figures will not match. Both are correct. They are answering slightly different questions.

The one-line answer. The dashboard tells you what is true now. The analysis tells you what the picture looked like when it was written, which is what its commentary is talking about.

Why we do not re-write past analyses

It would be easy to regenerate every report each night so it always matched. We deliberately do not, for two reasons.

A record that keeps changing is not a record. If you read your August analysis, made a decision from it, and came back in October to find it now says something different, you could not trust any of it. Keeping each one fixed means you can line up August against July and know that neither has moved underneath you.

The words and the numbers have to stay together. The commentary in an analysis explains a specific set of figures. "Revenue up 19% on last week, driven by higher appointment volume" only makes sense alongside the numbers it was written from. Updating the figures but not the narrative would produce something worse than a slightly out-of-date report: a confidently written explanation of numbers that are no longer there.

What actually causes the difference

1. Information that arrives after the analysis is written

This is much the most common cause. Your analysis for a week is written shortly after that week ends, and plenty of things about a week are still settling at that point.

Invoices raised or paid days later.
Appointments not marked as completed until the following week.
Clients booked in for a future session after the analysis had already run.

Two figures are affected by this often enough to be worth naming, and you will see a short note on the relevant cards inside the analysis itself.

Revenue received will usually read a little lower in the analysis than on your dashboard. Payments for a week keep landing for some time afterwards, and the analysis only ever sees the ones that had arrived when it ran.

Clients at risk of dropout will usually read a little higher in the analysis. That count is clients who have attended recently with nothing booked forward. Some of them book in during the days after the analysis is written, which quietly moves them out of the count on your dashboard but not in the saved report.

You will see this flagged. Cards for figures that are still settling when an analysis is written carry a short note saying so. It is not a warning that the number is wrong. It is telling you that this particular figure is the one most likely to have moved since.

2. Changes made in your practice management system afterwards

If someone deletes or corrects an invoice, changes an appointment's status, or tidies up a client record after the analysis has run, your dashboards pick that up at the next sync. The saved analysis does not.

This is worth knowing about mostly because end-of-month tidying tends to happen in batches, so a monthly analysis can be written a day or two before a round of corrections lands.

3. Changes to your own GrowInsight settings

This one surprises people, so it is worth spelling out.

Your status mapping decides which of your practice management system's appointment statuses count as Completed and which count as Cancelled. That mapping is applied across your whole history, not just from the day you set it. So if you add "Did not attend" to your Cancelled list next week, your cancellation rate for last March changes on your dashboards too.

An analysis written before that change keeps the rates it was written with. Nothing is broken. Your dashboards are now using your new definition everywhere, and the analysis is still using the old one.

The same applies to other settings that shape your figures, such as nominating an intake calendar, or changing a practitioner's configured capacity, which moves your utilisation.

4. How utilisation is measured

Your capacity is set per week, not per day, so utilisation can only be measured a whole week at a time. Months rarely start on a Monday, so each monthly analysis counts a week in the month it ends in. August's analysis, for example, might use the five weeks from Monday 27 July to Sunday 30 August, and the week starting Monday 31 August then belongs to September.

That keeps every week complete and counts each one exactly once. It also means utilisation, available appointments and unfilled appointments in a monthly analysis cover slightly different dates from the calendar month. A note next to those figures in each monthly analysis shows exactly which weeks it used.

Your Performance dashboard works differently. When you set it to a date range such as 1 to 31 August, it includes every week that range touches, including the part weeks at either end. So for a single month the two figures won't match exactly.

Some months contain four of these weeks and some contain five, so available and unfilled appointments can jump from one month to the next without your capacity changing at all. Utilisation is a percentage, so it's the better figure for comparing months.

Your weekly analysis isn't affected. A week is a week, and it lines up closely with the same week on your dashboard.

Only practitioners with capacity set are counted. Utilisation compares completed appointments with the capacity you've set, so a practitioner without capacity can't be included. If some of your practitioners don't have capacity set yet, your analysis says so, for example "covers 5 of your 8 practitioners", and you can add the rest in the Availability Hub.

Sometimes a utilisation figure isn't shown. If more appointments were completed than the capacity set for that period, the figure can't be right, so we leave it out rather than show something like 140%. A note beside it says which period is affected. It almost always means capacity needs updating in the Availability Hub, for example after a practitioner increased their hours.

Which one should I go by?

Every metric card in an analysis links straight through to the matching dashboard, with the dates already set to the period the analysis covers. That is the quickest way to see the current version of any figure you are looking at.

When a difference is worth telling us about

A few per cent on revenue received, or a handful of clients on a count, is ordinary settling and nothing to act on.

Worth flagging to us:

A figure that is dramatically different from your dashboard, not slightly.
A number you know to be wrong in both places.
A card marked Data quality issue, if you cannot see why it would be.
Commentary that describes something that plainly did not happen at your practice.

There is a Something look wrong? Tell us link at the foot of every analysis. It tells us which report you were reading, which saves you describing it.

Common questions

Can I regenerate an analysis so it uses current data?

No, and it is worth understanding why not. Each analysis is a fixed record of a period, which is what lets you read a run of them in order and trust the comparison. If you want current figures for a past period, open the dashboard link on any metric card. It takes you there with the dates already set.

Why does my weekly analysis sometimes take a couple of days to appear?

We wait until your data for the week has come across before writing anything, rather than analysing a week that has only half arrived. In practice that is usually within a day or two of the week ending, though it can occasionally be longer.

What does "Partial data" mean on a report?

It appears on a monthly analysis that covers only part of the month rather than the whole of it. The figures are correct for the days it covers, but totals such as revenue and completed appointments will be lower than a full month would show. That means comparisons with a previous full month are not like for like, so a drop against last month may simply be the missing days rather than a real decline.

Why is my analysis from six months ago showing a different cancellation rate to my dashboard?

Almost always because your status mapping has changed since. See "Changes to your own GrowInsight settings" above. If you have not changed it, tell us and we will take a look.

Why doesn't utilisation in my monthly analysis match my Performance dashboard?

Because capacity is set per week, a monthly analysis measures utilisation over the whole weeks that end in that month, while the dashboard counts every week your chosen dates touch. See "How utilisation is measured" above. The note beside those figures in each monthly analysis shows which weeks it covers.

How far back does my analysis history go?

Your last twelve months.

Does this mean the numbers in my analysis are unreliable?

No. Every figure in an analysis is a correct reading of your practice at the moment it was written. The difference you are seeing is the gap between a photograph and a live view, not an error in either.

Analyses are generated using AI and can contain errors. Use them alongside your own judgement when making decisions about your practice.

Page updated

Report abuse