Get Better Reporting: Put the “Intelligence” Back Into Your BI Solution
Does your current business intelligence (BI) solution really support your business processes by enabling you and your team to draw intelligence and actionable insights from marketing and sales activity? For most organizations the answer is “No”, making the “BI” moniker a bit ironic. Most organizations struggle with reporting typically because, the reporting solution is hard to use, out of date, or just full of too much data and clutter that requires lots of manipulation to be useful. This makes answering the “so what?” question very hard. There is a better way – here are a few quick tips to get more from your reporting:
How To Report on Marketing Tactics
1. Visualize the data
What would you rather start with to evaluate your business based on the options below? Would you rather have the chart on the right or the left (“Exhibit A” or “Exhibit B”)? The answer is obviously “Exhibit B” but there are a lot of “Exhibit A” offenders out there. Sure, we want to get access to the details but let’s not start there… go there only if needed based on a quick sub-second evaluation of the visualization that something might be amiss.
So let’s pull that thread a bit … in looking at the chart on the right (Exhibit B) it looks like our leads are steady and rising but we have a dip in converted customers as well as active customers.
As we dig into the customer data we recognize the importance of the next tip….
2. Benchmark the data.
While the raw values of things are important, it is critical to benchmark too. This might be a comparison to a goal, average, hypothesis, historic performance, or an industry standard. Look at the comparison of customers for the current and prior year below; without the prior year benchmark we would miss the seasonality (and possibly just declare our Q4 initiatives a total failure). However, relative to the declines seen last year, our efforts this year were able to reduce the percent reduction. We would never see that if we didn’t have the data or the picture on the right.
3. Keep it simple.
Users can only consume 5 to 7 metrics at one time, so trying to look at much more than that is just adding clutter. In addition, ensure that you differentiate between your KPIs and other metrics. Your key performance indicators (KPI’s) are just that – the key things to focus on as you evaluate the performance of the business. They should be afforded primary real estate in your reports and the focus of your conclusions.
4. Know your audience.
The level of reporting for an executive compared to a sales or marketing manager or an analyst can and should vary as the focus of knowledge shifts from “what” to “why”, respectively. Your dashboards in particular should communicate the information concisely and should not be confused for an ad-hoc environment with so many ways to slice and dice data that the user is overwhelmed. If you are building reports, keep this in mind. If you’re asking for them make it clear what level you need and what business questions you’re trying to answer.
5. Keep it moving.
Reporting should be agile with the ability to shift and evolve with the business. Expecting and planning for this is critical (you can’t set it and forget it with your BI solution). If your solution isn’t agile or user adoption is poor, take a look at the tool you’re using, the staffing you have, how people consume reports, and how data sets are integrated together behind the scenes. One or all of these is likely to be constraining you and the solution to fix it might not be as difficult as you think.
Lastly, while 5 tips would be nice and tidy, I can’t ignore one of the most important points:
6. Don’t lie (or be fooled!) with statistics.
Inspired by one of my favorite books How to Lie with Statistics, there is a lot to consider here. Be mindful of how you’re visually communicating (or consuming) the relative scale of things, how things are (or are not!) benchmarked, whether a metric is truly meaningful (or statistically significant) and the level of precision that is or is not present.
Take for example, the simple line chart from earlier… I can simply stretch it out if I’m trying to communicate that things are really quite steady.
Or consider the first data point (January) on the comparison chart of the current year versus the prior year. While the number of customers increased by 11% the wrong visual can make the improvement look much larger (Note the y-axis on the chart below and its misleading scale). I didn’t manipulate this one either, it was an auto-axis chosen by the tool. Our data tables can be misleading too. For example, just because a metric is or can be carried out to multiple decimal places does not mean that comparison at that level is valid. For example, if we convert 4.299% of leads this month versus 4.225% last month is that difference meaningful? Are we doing “better”?? The fact that the metric wasn’t rounded to 4% would indicate that maybe it is but the samples we’d need to be working with to make this a statistically meaningful difference (or even a valid directional one!) are enormous.
As consumers and builders of reports, there are a lot of opportunities to save time and get more intelligent by being mindful of some of these tips. Happy reporting!
Want more analytics tips and strategies? Download our whitepaper How to Leverage Predictive Analytics.