Digging into your data

Data visualisation has come a long way in the past 25 years since I first made a chart in an excel spreadsheet. There are a number of tools available to help you to visualise your data or to do it for you like Actually Data Analytics ADRFM suite. The challenge remains though that if you don’t know how to read, interrogate and understand your data then you’re not maximising it’s potential.

My 5 year old can look at a bar graph and say which number is the highest or look at a line chart and tell me that the line is going up or down, but to make these charts truly useful we need to be able to dig into the reasons behind these numbers. Once we understand the reason for certain trends we can then use this information to make informed decisions about how best to use budgets to maximise the desired results in the future.

Some fundraisers and Database managers will be able to tell you fairly accurately the numbers which a chart will show, before you’ve created the report, but even these people still find some interesting revelations when digging into their data using ADRFM.

Here is an example report showing the number of individual payments per quarter.

The scale is altered to start at 36,000 which over emphasises the difference between the columns.

We know that more payments were made in Q2 but without more detail as to why that’s not really that useful a piece of information.

Here is the same chart but broken down into two levels. Here group Y are Regular payments and group X are One off payments.

Now we can see that the Blue columns for group X have an increase in Q2 and group Y increase in Q2 & Q3.

This could be a seasonal effect with more annual donations during the summer or Christmas period or due to a particular campaign or event maybe. To find out more we’ll have to dig a little further still.

Here is our chart one final time, this time we’ve broken down the columns by ACORN categories. Now we can see that the increase in Q2 is largely driven by people whose postcodes are in the Rural Residents group. So we can use this information to target these areas again in future campaigns.

The next steps would involve looking at each of these groups in detail to understand if particular campaign were effective/ineffective, to see which payment methods and channels each group favour and how likely are they to make another payment. Is that second gift to the same type of campaign & product or are they susceptible to other products and opportunities?

As a data geek, I could spend hours (and on occasions have), looking into various combinations to look for trends and patterns, hidden golden nuggets of detail which the database will give up to me. This research can be invaluable, but there is always a risk of going down a rabbit hole and losing hours to find so much detail that the marketing team will need to segment the next mailing 250 different ways to make the most of it, which in most cases isn’t practical.

Of course there are a number of different ways we could have looked at the data. Age and Gender have in the past been the most common ways of breaking data down. Although base Genders can be derived from most titles it doesn’t always help and targeting genders is not as straightforward as it once was perceived. Similarly ages are often looked at, but in most databases less than 5% of contacts have supplied their age or D.O.B. so the sample can be misleading.

Most contacts will have an address, so postcode data can be used for a good sample set. There will always be some anomalies when generalisations like this are used, but it’s a fairly solid platform to start from.

Tools like ADRFM, when used correctly, can help us to quickly identify groups and trends which we can practically use for future decision making, making our communications more effective and helping to maximise the return on our spending.