Behavioural Data patterns: The synergies between mass media, social media and mobile media

We are seeing the emergence of three distinct media types: Mass media(TV), Social media(Social networks) and Mobile media.

Often these media types are discussed in isolation – but they make a lot more sense when viewed holistically in a converged ecosystem from their underlying data patterns

The idea is: Spot the patterns in data usage and validate the patterns using promotions. Do this over as wide a network as possible(either a converged network or in partnership with a broadcast media channel)

The goal of any media is to get the receiver of the message to take action – ideally a purchase decision(but not necessarily so)

Broadcast media/ Mass media(TV, newspapers etc) are great for brand building. They are best placed to get the message across to the widest section of the public. One could argue that mass media is ‘interruptive’ in nature – but without an interruptive message – I can only know what I already know – but I cannot know what I do not know ..(At the risk of sounding Rumsfeldesque :) .. ). On the other hand, a message that can be personalised to the recipient is truly useful to them. So, the goal is to work with these two end points

Also, irrespective of how the message is received, ideally the two ends of the media cycle need to co-relate i.e. for every recipient of the message – there should be a corresponding action. That action should be tangible and measurable in some way – i.e. you need to be able to measure how many people received your message and how many of these people took action based on that message.

Traditional media focussed on these two ends of the equation.

traditional media.JPG

Into this mix comes social media and mobility – both of which have a unique and a disruptive effect. What I am trying to show below is that they have a synergistic effect when viewed together:

The effect of social media is as follows:

The defining characteristic of social media is it’s two way nature i.e. social media participants are not passive consumers of the message but also potentially creators of the message and indeed have the ability to mutate that message. So, social media could be the reinforcer of the message or it could be the destructor of the message. Hence, on one hand we get the viral effect (reinforcer) and on the other hand we get the death spiral for the message (destructor)

social media.JPG

The effect of mobility is also very disruptive ..

The mobile operator /provider has the opposite problem from the broadcaster. They know the customer individually but do not know much about that customer. Specifically they do not know the user’s preferences which can be used to create a more personalised message (and by extension a message that is potentially useful to the receiver)

So, the approach could be: Spot the patterns in data usage and validate the patterns using promotions. Do this over as wide a network as possible(either a converged network or in partnership with a broadcast media channel)

mobile media.JPG

Mobile media complements mass media (like television) by providing the individual personalization which mass media lacks. This can happen in potentially three ways – all these ways are interconnected

a) Patterns: Spotting behavioural patterns

b) Promotions: Specific promotions to validate observations from the patterns

c) Partnerships: Creating partnerships between traditional media and mobile media to get as wide a network as possible(or extend your own network in a converged ecosystem)

Patterns: Patterns in usage can be gleaned by merging data streams from outside the telecoms network along with Telecoms data (CDRs).

Consider program viewing schedules. Suppose an Operator analyses their data based on program viewing schedules and discovers that at a certain time every week, a group of people on their network communicate via text message.

If that time co-incides with the transmission time of a specific program(say a TV serial) – then potentially these customers are fans of that serial and are communicating with each other as they watch the program.

This information can be validated via promotions i.e. specific promotions designed to test the hypothesis with an opt in clause for further promotions.

Finally, a telecoms provider can get insights if it partners with a TV station. They can also extend the scope of the network itself beyond mobile. This can be done by offering fixed network services, IPTV services etc – either on their own or in partnership.

The key is to get a single view of the customer from their data patterns and / or to get an incrementally detailed view of their customer through their social interactions(who they interact with) and behavioural interactions(what type of behaviour do they exhibit – which may need external data sources) – and then refine that analysis by specific promotions to validate the observations.

The wider the network they can do this over(either on their own or in partnership) – the more effective the result

Many thanks to Tomi Ahonen for his feedback on the mobile related ideas in this blog


  1. jMac says:

    Some very good insights here.
    I don’t tend to split social media from mobility (although I can see exactly why you have above) and as for ‘social networks’, personally I view them as the connectivity between our social group. Social networks existed long before ‘Facebook’as we know….and I mean by thousands of years.
    I agree entirely with the use of interactive channels to complement ‘make interactive’ the monologue channels.
    Your article raises another point about mobile operators knowing who the customer is.
    We should now start a movement of transparency and reality on this point.
    1. Operators know who the billing point is registered to be – so long as its a contract play. They cannot 100% identify the customer, only the data stated on the bill.
    2. If it isn’t a contract play, operators only have behaviour to go on – from SMSC/Click-stream etc – and at this juncture it is worth pointing out that the trend toward pre-pay in some territories completely ECLIPSES contract play. What then for connecting the unknown with brands?
    3. Al operators have a significant inaccuracy in their data sets. Let’s be honest about this. Even on the accurate data, the silos are often not joined together…..after all, the model was/is based on billing people rather than learning anything about them.
    4. Almost all of the information they have does not match the information advertisers want and need. Incredibly this isn’t front of mind just yet in wondering why the ‘year of mobile’ hasn’t happened yet. The answers are clear in my opinion…
    5. I believe there needs to be at least 10k segments created so as to ensure minimal risk of irrelevance. So far, most major operators can present inventory in around a dozen, which is incredible if applied to tens of millions of people.
    6. I see another 3 P’s as PERMISSION, PREFERENCE and PRIVACY (as written here: I believe the mobile opportunity cannot grow as much as we expect/would like without peoples permission and preferences known by ASKING rather than inferering or assuming.
    7. Patterns (in your article) is a good approach BUT, the ‘gleaning’ and aligorythmic practice only reduces failure rates by points rather than moving to abolishing them by flipping the model (as in point 6).
    So – in summary – great post Ajit and I like the approach of writing guidance in how we can best use this unique and ubiquitous channel. Adding some home-truths to it will justify the points even more and – I believe – add even more clarity to the way forward. There is no way back from here.

  2. Ajit Jaokar says:

    Many thanks Jonathan! appretiate your detailed insights! kind rgds Ajit