Getting more from your research stats

Media and brand strategists are expected to have all the answers, and with research tools like AMPS at their disposal, they could, in theory, meet that expectation. There are huge quantities of data across the AMPS, RAMS, and TAMS industry research which we all fund. Two problems however, are standing in the way of people's getting the most out of this research - a lack of time, and a lack of knowledge.

If I were to fully interrogate AMPS, it would take my entire working day, week in and week out, because there is just so much information in the database. With crushing time pressures however, most of us only ever find time to look at demographics, LSMs, and AMPS, RAMS and TAMS media information, throwing in a look at the branded data if the brand we're working on is featured.

For many, not drilling down into the data is a simple case of running out of time. However, more and more, it's also because people don't have the knowledge of what is contained within the industry research, and what they can do with it.

Understanding data

The industry has lost its statisticians, those great data analysts who used to have their own departments, and who really understood data, and knew how to read and interpret it. A lot of analysis demands a statistical background, but there is little statistical training in the industry nowadays.

As we go forward, it's going to get worse - static bottom lines and squeezed margins will mean less staff, and the lack of training and understanding will compound.

This is one of the reasons why people cling so doggedly to packaged tools like the SU-LSM, and the preset age groups and light, medium and heavy usage definitions.

Not many people realise, for example, that you can create your own data codes as a filter for future analyses. When working on a beer brand, we once recoded the definitions for light, medium and heavy beer users.

By default, heavy beer usage is five or more beers a week, but that's not even a beer day. We wanted to be able to look at people with far higher consumption, so we reclassified light, medium and heavy. Branded AMPS lets you run exactly how many people drink from one to 50 beers a week, by brand. If you look at this detailed data, you can more easily pick up a middle point, which gives you a more accurate idea of what is heavy, medium and light usage.

Most people only use what they can see, what they're given by the programme. Most of them don't know how to change the presets, or even that they can be changed. Yes, it takes time to code the data, but once you've coded and saved it, it's really easy.

With a little creativity, you can really make the AMPS data work for you. In conjunction with Peter Langschmidt, we developed a model based on AMPS to help us put people into electricity consumption groups, for Eskom's communication plan. We used LSMs, because we needed to work out how many kilowatts various households were consuming. Therefore, we'd see which households had dishwashers, for instance, and then overlay how many kilowatts would be used by this appliance on a monthly basis. We can then judge, by SU-LSM, which are the most electrically expensive households, and can target specific groups of households with tailored energy-saving messages.

This is data which we, as an industry, have commissioned and paid for, so we might as well get the most out of it. And, there really is an enormous amount of valuable information which you can mine, with a bit of time, and a good measure of knowledge.