You walk into a meeting and find that everyone has brought reams of reports, dashboards, graphs, and pie charts; that everyone’s reports reach a different conclusion; and that by the end of meeting you can’t even recall the problem you were trying to solve.
If this sounds familiar, you’re not alone. Data, Analytics, BI tools have become democratised - they are now available to more people in more departments across an organisation, and are much easier to work with. All this is good news - but it can also create headaches. So here are a few rules-of-thumb to help you reclaim your sanity if you’re drowning in reports but gasping for clarity.
Definitions (or how to get your stories straight)
If you ask 5 random people in your company what a DAU (Daily Active User ) is, will you get the same answer ? Will they all list the same set of activities that a customer must do to be counted as active ? What about Sales ? Does everyone agree that a sale happens when the product is shipped ? Or is it when the payment is received ? Or maybe when the refund-period has expired ?
The point in all this is that it does not even matter which exact definition you use - each definition in itself is logical and gives you some nuance of information. The trick is to have everyone use the same definition so that everyone knows what anyone is talking about.
Make sure definitions are agreed, approved and in a change control system in your company. This is something that is often considered trivial and delegated to a junior team who struggle to get agreement across the organisation. Instead, give it some management time in the beginning, and it will pay dividends in clarity for months and years.
Looking ahead vs Looking back
Typically you will get a lot of reports and dashboards in you inbox which will be looking back - what happened yesterday, what happened last month, and so on. Accounting for the past is great, but having ten people tell you last month’s sales were missed is not helpful as having two people suggest that next month’s sales are likely to be missed.
So prioritise reports that try to tell you what is about to happen. If you are not getting enough of these you’re missing out.
Understand that forecasts are not prophecies
Related to the last point, one of the reasons there are fewer lookahead reports is that people are afraid of being wrong. That’s because a lot of managers and executives don’t understand that forecasts are just that: forecasts.
So create an environment where everyone understands that forecasts are statistical animals and should be used a jumping-off point to plan ahead. Let people bring their judgements and theories into the forecasting process, use these to brainstorm, understand, and plan. Encourage people to have a view about the future based on their analysis.
Demanding that forecasts be prophecies will only mean that you get fewer and fewer lookahead reports and more and more accounting of the past.
Internal vs external data
From accounting to marketing to operations, everyone will have reams of reports based on your operational data. But make sure you are also looking at data that comes in from the outside in: your market research, your direct customer feedback.
Ideally you want to have the same set of people doing both types of reporting. If your DAU are going down, your internal data will tell you a lot about which customers have become inactive, what were they doing before leaving, and so on - but also complement this with reaching out to some of those customers via market research and asking them what happened.
No amount of correlation analysis can match simply asking even fifty customer directly: what went wrong ?
Have a cadence for your reports
Simply getting tons and tons of reports at all hours is overwhelming. Try to corral these into some semblance of order. Here’s my system: I like to have daily reports which are more operational, weekly reports which are trends and lookahead based, and monthly reports which are deep-dives into specific issues.
For instance if you have five daily reports coming in at all hours - each with an overlapping set of recipients, try to have a single daily report that contains all the information. One thing I've noticed is that people tend to try and make their daily reports no more than one page - some of them are even called one-page-reports. But this is just a mental block, specially now that most reports are online and viewed on different screen sizes anyway. Having a single 5 page report every day is infinitely better than having 10 one-page-reports.
Together with the definitions, the cadence will ensure that everyone in the organisation is aware of the key facts and forecasts of the operation and can count on reaching out to others with the confidence that they will have seen the same numbers.
Let a thousand flowers bloom
None of what’s said here should imply that reporting should be the domain of a small set of specialised people with everyone else only a consumer of information.
On the contrary once you’ve set down the definitions, created a cadence of core reporting workflows, give everyone access to all data points. Let them experiment with it, understand it, forecast on it, learn insights from it. The more the number of people who use, manipulate, and correlate data first-hand, the more organisational intelligence you will build - just make sure you have the basic rules of play in place to benefit from it.