StanChart’s price tracker rolled out in Nigeria is a great example of where and how mobile phones can really add value in understanding the African consumer market and add substantially to its scarce database. What concerns me however is the increasing promotion of the ubiquitous cellphone as the means to gather consumer insights for all sorts of polls, surveys and sentiments.
Surveys conducted online and through the phone may not, at this point in time, offer a representative sampling of the relevant population, no matter how random. Ironically, in this context, its this very randomness that creates skewed results. Unless the results and the methodology clearly specify the gender, age, income and education breakdown of those responding to their survey, there’s little basis to assume that they are representative of the population. Reliance on such results should very much be contextual – which country, what are they aiming to show, who exactly did they survey, rather than accepting results from any old location on face value.
Here are some recent stats that help explain why:
The Mobile Africa 2015 study, conducted from GeoPoll and World Wide Worx, surveyed five of Africa’s major markets; South Africa, Nigeria, Kenya, Ghana and Uganda finding that mobile Internet browsing now stands at 40% across these markets – Ghana: 51% Nigeria: 47% South Africa: 40% Kenya: 34% Uganda: 29%
And these are the top 5 markets.
Let’s say you get results via mobile surveys – you’ve already narrowed down your sampling base to less than one third of the population. If you’re not calling them up, then you’re narrowing it further than those who can read and write, and if your survey was in English or French, its narrowed further to those educated in the language. By the time you actually get to the people responding to the survey, you’ve effectively sampled a tiny unrepresentative slice of the national population.
If I wanted to know what young tech-savvy men think, I’d never hesitate to use the results of a mobile survey. If I wished to have a better idea of lower income or female heads of households, or even those in regional towns and cities, I would be sceptical of any research conducted without human intervention. There’s also a high risk of surveys being filled in for the nominal cash or equivalent rewards. There isn’t enough quality consumer research available on the African consumer market that we can risk further muddying the waters like this.
On the other hand, as this StanChart price tracking system shows, there’s a lot of untapped potential for the use of phones in consumer market research across the entire continent. It just may not necessarily be something that works in exactly the same way in the OECD world.