What it is and what it's for#
The programme leaves a record of everything that happens: who logs in, what they complete, what they redeem, what they answer. Analytics turns that record into two different things, and it is worth keeping them apart because they are fixed in different ways:
- Is the programme alive? Participation, completed challenges, streaks, use of points. If this goes badly, you fix it with content and communication.
- Is anything changing in the organisation? Climate, recognition, what people write. This moves slowly and is not fixed with more reminders.
There is no single dashboard that answers everything. There are several, each with a question of its own:
Engagement dashboard
How the programme is doing: the headline metrics, the breakdown by challenge and by location, and the named list behind each number.
The manager's panel
Who to call. The roster of their team, person by person.
Pulse and sentiment
What the mood is, by category and by location, and how it evolves.
Comments dashboard
Where the conversation is happening and what is waiting to be dealt with.
Survey results
What people answer, and the difference between anonymous and named results.
Smart recommendations
What the platform suggests you change, based on what it sees.
How to set it up#
The metrics calculate themselves: there are no reports to build and no metrics to declare.
The only part of analytics that you do configure is the pulse question catalogue, because there you decide what gets asked and in which categories. It is in Pulse and sentiment.
What you do decide on every dashboard is how you slice the data: the period and, depending on the dashboard, the location and the organisational unit.
The slice is what makes the figure useful. A global participation rate of 60% does not tell you what to do; 85% at some sites and 20% at others points exactly to where you should go.
Not every dashboard offers the same slices. Before comparing two figures that come from different screens, check that they are looking at the same period and the same set of people. Two similar numbers that do not match are usually two different questions, not a calculation error.
What happens next#
- 1
First check whether the programme is alive
The engagement dashboard, top to bottom. Without activity, the rest of the metrics mean nothing: a good climate measured on fifteen responses is not a good climate.
- 2
Then, where it gets stuck
Slice by location and organisational unit, and go down to the breakdown by challenge. The differences between sites say far more than the average.
- 3
Go down to the names
Almost every metric has a list of specific people behind it. That is where the figure turns into something you can act on: a call, a reminder, a visit.
- 4
Finally, what is changing
Pulse and surveys, comparing periods. Here one month's figure is worth nothing; the series is.
Frequently asked questions#
Which metric should I look at if I can only look at one?
Active people in the period, sliced by site. Almost every other problem shows up there sooner or later.
What counts as good participation?
It depends on the type of workforce and on whether the programme is voluntary. More important than the absolute number is the trend and the spread between sites: a stable, even 50% is a better sign than a 70% falling month on month.
Why does a journey's completion rate drop if nobody has stopped taking part?
Because you have added challenges. The numerator is unchanged and the denominator has grown. It is normal and it recovers.
Can I export the data?
Yes. On the engagement dashboard, the detail behind each metric downloads as a spreadsheet file, and you can also copy the email addresses of the people on that list in one go. Anonymous survey responses are exported without an author.
Can I see the detail for one specific person?
Yes, from their record: their progress, their points and their redemptions. Anonymous survey responses, no, by definition.
Two dashboards give me different figures for the same thing. Which do I believe?
Before assuming one of them is wrong, compare the period and the scope. The manager's panel looks only at their team; the engagement one, at the whole project. If they still do not match once period and scope are aligned, say so: that is the kind of discrepancy worth looking into.
Limits and warnings#
- Inactive people drop out of the lists. If you deactivate a lot of people, the percentages rise without anything having improved. It is the easiest way to believe in an improvement that does not exist.
- Points pending approval do not count as issued. A large approval queue distorts the reading of the points economy.
- Anonymous responses cannot be traced down to the person. They are read in aggregate; that is exactly the deal you made with whoever answers them.
- Comparing waves requires that the questions have not changed. If you edit them in between, the series breaks and stops being comparable.
- Data with few responses is not representative. Be sceptical of a category or a site with a handful of responses, however striking the result.
- Culture metrics need time. Drawing conclusions from a single survey wave, or from two weeks of pulse, leads to the wrong decisions. Always compare against the previous period.