Marketing opinion

Subscribe to industry newsletters

Press offices

Enquire about a press office
Bizcommunity has over 400 industry contributors and we always welcome further contributions and contributors.
Advertise with us
Advertise & RatesMy Account
Company press officeList company
Recruitment packagesSubmit job ad
Download ratecard

Marketing opinion

Segmentation models and why they often don't work

I'm always amazed whenever I see a client's segmentation model. It often feels like seeing a dinosaur fossil, because that's exactly what these models are to me. I've come across a company that routinely (about every six months or so) re-segments its market.
I thought that the idea was ludicrous, mainly due to the absurd amount of cost and time that go into these studies. Upon ruminating I realised that I find the entire idea of segmentation utterly offensive. Here is why.

Remnants of a bygone era

I think that I just don't get segmentation models because businesses tend to segment their markets based on the most crass and rudimentary data available to them. It's usually things like income, age and race. Of course these factors impact on the way people will respond to products and brands, but surely these measures are archaic in today's fast-paced, everything-at-my-fingertips society?

Removing the human element

Think of the person you love most in this world. Now think about into which of your segments that person falls.

Does the segment even begin to describe the person you thought of? Their complexity, emotions, thoughts, likes and dislikes? Probably not. And that is my major gripe with these marketing relics.

We seem to be so caught up in trying to place our consumers into convenient boxes that we forget that they're flesh and blood people.

Quite a few companies have realised that the traditional way of segmenting the market isn't the way to go. So they've tried attitudinal segmentation. Another fail.

Our individual attitudes and opinions are as unique as our fingerprints. How could we, as marketers, ever be as arrogant as to think that we could group people together based on their ratings of a few sentences?

We are all outliers

After analysing hundreds of data sets I'm pretty au fait with removing data outliers and analysing data within an acceptable variation of the mean. The principles of segmentation remove all variables that aren't within a miniscule percentage of the mean, and that is why I disagree with it.

We can't be everything to everyone

I believe that too many brands today attempt the one size fits all approach. Admittedly, many segmentation models are designed with the aim of countering that type of thinking, but it's also why many fail. Why try to talk to different consumers differently?

The simple fact is that people do not like being neatly packaged into generic little boxes. Once we start talking to people plainly and honestly, without all the marketing hogwash, we will gain their attention. That's a pretty good place to start if you ask me.
    
 

About Cyndi de Vries

Cyndi de Vries is a 25-year-old Research Consultant at Interact RDT, a customer experience agency based in Melrose Arch. She's been in the research industry for about four years now and can't think of anything she'd rather do. She is passionate about customer experience and helping organisations positively affect it.
Leon Jacobs
It is also true that for some sectors, like telecoms, it is impossible to segment all the possible combinations of need states and demographics.
Posted on 13 Oct 2012 12:44
James Gradidge
Why try to talk to different consumers differently?

Because that is one of the founding principles of modern communications - speak the language of your consumer and you are more likely to be heard.

Address your consumer from an ego -centric position and you are most likely to be ignored.

Talk the Talk, so you can Walk the Walk.

Consider a category such as "music lovers". How many genres, sections and sub-sections of sub-sections are represented? In fact one group may be so far removed from the other demographically, that they would not even understand each other. What can you possibly say to them that they would all understand?

Even music, as the"universal language " has it's limitations, for this very reason.

More important though, is the necessity to locate the leaders in each segment and understand their choices and motivations, before you say anything. Then you can start tempting them to become followers. The rest will follow, naturally.

Quantified data, no matter how detailed, can only go so far in defining the path of your communications.

The rest is art, experience, and dare I say it, genius.
Posted on 14 Oct 2012 02:31
Gavin Taylor
While I see the point you are trying to raise, your viewpoint seems to be a little biased to the world you experience daily. You must remember, that although the world is 'Everything-at-my-fingertips', most of South Africa is not. only 10% of this country is on Facebook, and far less are registered on Twitter. 9% of this country have purchased something online in the past 12 months. that leaves an astonishing 90% of South Africans who are very much still living in the bygone era where segmenting the population by demographics is the most effective way of reaching them

Just my 2c...
Posted on 15 Oct 2012 12:39
Cyndi De Vries
Good point well made Gavin. Thing is, although 88% of the population are not online, I still think that there are better ways of reaching these consumers.
It's the quantitative approach to segmentation that is my real gripe. Why not adapt the principles of ethnography in order to learn how to speak to consumers? There are so many other avenues of exploring customer profiles, but we seem to be stuck on this very rudimentary way of segmenting markets.
Posted on 15 Oct 2012 13:41
Gavin Taylor
Now you are speaking a language I can agree with! Ethnographic research is still very new to South African marketers... I too am in research, and every single time we pitch an ethnographic approach to a potential client, the reaction always seems to be along similar lines: 'How many people is that relative to and how do we quantify that?' (in a nutshell). I agree that we need to be constantly re-inventing ourselves, but that, by-and-large, starts with the multinationals... Marketing Managers have been so overwhelmed with quantifying everything that even Facebook (however significant in the marketing context it may be) has come down to the numbers, when it should be about the conversation.
Demographic segmentation seems to be the path most travelled and in our current economic climate, not too many corporates are keen to get onto the 'Lets Try Something New' bandwagon..
Posted on 15 Oct 2012 15:29
Bob Lewis
Your problem seems to be about demographic segmentation - which is only one of many models for segmentation - and the most basic.

You should perhaps know that over time, there have been many highly sophisticated models constructed via intensive research that break through the simplistic-ness of race, income or other demographics. For example, the Heylen Implicit Model quite convincingly segments all humans into 8 basic fundamental psychological typologies, which can be demonstrated to govern their attitudes and purchasing behaviours across scores of product categories.

Brilliant work - but - which like all similar models - has the basic weakness for marketers of being significantly un-exploitable when it comes to most communications channels. Most psychological typologies are not precisely leverageable in the majority of media channels- with certain exceptions like special interests - which is why I believe that special interest magazines supported by websites are the singularly most powerful media combination currently around (SA has the problem of a very small affluent market base in this respect.)

You cannot optimally market one utility to 8 different typologies within a one brand business - you cannot afford an 8 brand-positioned business selling the same utility, either.
What you inevitably end up with is a collective and thus compromised, brand proposition in overlapping media.

Funnily enough after 30 years in the business my observation is that the basic demographics you despise, have enough behavioural and attitudinal correlations to make them practically applicable in the majority of categories. You do require a seasoned perspective and interpretation of the opportunities, however.
Posted on 15 Oct 2012 16:20
merlin naicker
Very interesting views and opinions. I agree with sentiments in so far as the reliance on demographic data is concerned. I have been in the media space for the last 18 years, and have long held the opinion that demographic segmentation will just cut it any longer, as the market evolves. Sometimes faster then the products servicing it.

Media is driven by numbers, i.e audience ratings, box office sales, CPP's, etc. This in delivered to media buyers and planners to achieve their targeted reaches. But along the way we loose some vital information, which has permeated in this "NEW MEDIA" platforms. I believe that the traditional segmentation (Tried and Trusted) approach has not delivered the proper yield and markers have rather tried to make it fit as opposed to seeking new models. Case in point would be the large number of facebook accounts, but no real way to monetise it....

As the world converges, the traditional approaches are unable to explain the phenomenons that are being presented. We need to be looking at our markets in a different way, that does not rely on stereotypes, and LSM levels.
Posted on 16 Oct 2012 15:38
LEGAL DISCLAIMER: This Message Board accepts no liability of legal consequences that arise from the Message Boards (e.g. defamation, slander, or other such crimes). All posted messages are the sole property of their respective authors. The maintainer does retain the right to remove any message posts for whatever reasons. People that post messages to this forum are not to libel/slander nor in any other way depict a company, entity, individual(s), or service in a false light; should they do so, the legal consequences are theirs alone. Bizcommunity.com will disclose authors' IP addresses to authorities if compelled to do so by a court of law.

News