Alan and I have been discussing the role of social context in mobile advertising. Recently Arun Sarin, the CEO of Vodafone spoke about the increasing role of mobile advertising, saying that mobile networks know who we are, what age we are, where we are, etc. This is all of course central to understanding mobile as the 7th mass media, radically different from the internet and TV, in fact far more far-reaching and powerful as a mass media.
I think its time to explore this area. There are three distinct concepts. I'll call them our identity, our digital footprint, and the social context.
IDENTITY IS USELESS
Our identity is the real honest facts about who we are. Not what we do, or with whom, but only who we are. Our name, our home address, our age, marital status, our employment info, our social security number, passport number, our face, our fingerprints, eye scans. Who we are.
This is often the aim of various marketing efforts, to try to get more accurately that information about customers. Where do we live - marketers like to think that our home zip code has relevance in marketing - the uber-rich may live in Kensington in London or Mid-town Manhattan in New York. But this pursuit is a fool's errand in today's digital world, in the time of mobile phones.
First, it makes no difference to you attempting to sell a product to me, if you know my name as being Tomi or Thomas or Tommy or Tom or perhaps its actually Larry or Barry or Gary. Maybe its Tammy - as in a woman. And come to think of it, Tomi could just as well be a woman as a man. What do you really know about me if you know my name. Nothing.
Yes, it is possible to customize communications coming to me "Dear Tomi.. or Dear Mr Ahonen.." but the novelty of that gimmick ran out in the 1980s with junk mail. Nobody today is impressed if an automated emailer gets my name right. No value in knowing my name.
The same is true of all classic identity information. Take the address. Yes, in extreme cases there is some value. If you are selling Rolls Royces or Aston Martins, then yes, if the address is South Bronx (the most poor parts of what is commonly thought of as New York's black "Harlem" area and beyond) - then yes, that is unlikely - but not impossible - to get Rolls Royce customers from there. Now, realistically, there still are the occasional drug dealers and pimps who might be in the market, but yes, I understand in extreme cases we can get some value from the address.
But lets look at it from the other side. How about a rich neighborhood. So the person has an address in Kensington in London or say on the lower East Side of Manhattan. The really posh addresses. Yes, that person might be rich with that address. But the address is not infallible, it could also be the butler or maid working for a rich person. Again, just having that address is no guarantee it is the type of person we'd hope.
What is more, the address information is often out of date, and even manipulated by people. For example when I left home at age 18, I moved in to live with my girl friend of the time. That was not a decision taken at one sitting. It happened very gradually. First I'd spend a weekend at her place, then a given night during the week, and then a week went by that I didn't go home, soon it was several weeks in a row. Then my parents asked me, Tomi are you still living here - to which I still replied - yes of course I do - although for all practical purposes I lived at my girl friends's apartment in another part of town and spent two or three nights per MONTH at home anymore.
By every measure of formal address, I'd still be registered with the same address as my parents, but in reality I had already moved away. If we don't ask our customer what is the address but rather we observe the behaviour of the mobile phone, we see exactly where this young adult is spending all his nights...
The same is true of all other demographic data, our marital status, age, etc. Useless (when compared with digital footprint and social context)
Digital identity gives us no actionable information. If we use this type of data, we get false positives (a person seems like being rich, is not) and we get false negatives (we assume a mid-market address in Queens in New York is not rich enough for our car, but the owner there has lived in the family home for five generations and is filthy rich).
Demographic data on prospective customers, was better than nothing, in the last century. It is worth throwing into the garbage bin today. We have FAR more powerful and accurate means at our disposal. They are on mobile. So, lets leave the identity, and move to the digital footprint.
By digital footprint, I am not so interested in the full knowledge of the "who" but rather the "what." What is it that I do, what is it that I consume, and where and when. Lets start online. On the internet we leave digital footprints, what we consumed and when. But the information collected on the internet is very weak, because of firewalls, shared computers, multiple accounts, anonymous accounts, deliberately mis-directed information (I am a 101 year old Alaska resident often on web surveys) and the fact we delete our cookies. The internet had the potential to collect powerful actionable information, but today that tends only to be within given closed environments, like on Amazon. The Amazon recommendation engine is remarkably accurate, and a sign of where we can go.
But mobile takes this into a whole new dimension; a whole new galaxy above and beyond Amazon. On mobile we know so much about you, the end-user.
Take a typical London resident, working in the City. He (or she, we don't care at this point) spends most weeknights and weekend nights at one area of London, regularly. Lets say the Earl's Court area. If this customer spends most nights there, that is the real home, the real address. While we don't know necessarily the street address of this guy, we know at least by cell identity, which section of Earl's Court the person lives in.
Then lets take phone movements. Is the person in a regular job (and/or school). Leaves every weekday morning with the phone moving on a near-identical pattern. But as it leaves the home, it sometimes stops at Earl's Court for short periods of time, sometimes 2 minutes, sometimes 10 minutes, sometimes not at all. Coming home, he almost never stops at the same location. A-ha. This is most likely a commuter using a bus. Going to work, he walks to the nearest bus stop, takes a London bus to work. Sometimes has to wait a little while, sometimes not. Coming home, no reason to stop at this bus stop to wait, goes straight from the bus onto home (or local shopping etc).
Ok, we have now found the pattern, this is a commuter, going to work (or school).
And of interests. Lets track the person for a few months and see what kind of special behaviour the phone movements indicate. The person went to Heathrow airport, then shut the phone off, and it was turned back on three days later at Heathrow. Very clearly this person took a short trip somewhere (but didnt' turn this phone on at the destination country). The person probably likes to travel if that was a weekend, or perhaps travelled for work if it was during the week. We'll monitor that more.
Then how about a special occasion. The same person's phone shows up at Silverstone (Grand Prix car circuit) on the Grand Prix weekend in July. Oh, this person is a Formula One car racing fan. Or if the phone spends several hours in Wimbledon during the tennis championships, we have a tennis fan. And so forth. Not yet absolute certainty, but the digital footprints are giving a powerful indicator of who it is and what they do.
We leave digital footprints to the mobile network and these are all actionable. We don't share our phone, and we take the phone everywhere. Even if we have two or more phones, the network still knows that every time this one phone number 0123456789 does something, it is always the same person, and that person is very certainly distinct from another phone number like 0987654321.
That was all just movement, no calls, no messages, no internet access, no mobile commerce. Now lets move to today's world. Our user accesses the bus schedule service for real time information on a given bus line. Ha, we know more. That same person uses the mobile phone to pay for the London congestion charge this Saturday and again next Thursday - yes, the person pays for the London congestion charge, is very likely the car owner, not only a passenger in the car. Then the person buys a Pepsi at a vending machine and pays by mobile. Over the past month we find eight times bought a Pepsi, twice a Coke, once a Sprite. This person has a preference for Pepsi rather than Coke, and we know its not a "diet" drinker, none of the 11 drinks were the diet variety.
The digital footprints start to collect. The person downloads a raunchy game onto the phone, a game with semi-nude women. Now we know the person likes naked women (as opposed to liking naked men). The digital footprints collect. The more the network allows us to consume content - Google something, see a WAP page on something, subscribe to SMS alerts on something, we get real, honest, actionable data on this customer. The true digital footprints, more honest than anything the person might have admitted in real life on any surveys.
And as we add the interactive elements of marketing and advertising, that are run through the phone, we get much more insights into this person. So they run the latest version of Big Brother or X-Factor (UK equivalent of American Idol) on British TV. 20% of British TV viewers vote in these kinds of shows. Does our target vote on an interactive show? Yes? We now know something about this person's TV habits. Not yet everything, but again more. The digital footprints collect.
So we run a campaign with Adidas to win some shoes. This person responds and we now know the person has an interest involving running shoes. But more importantly, if we structure our campaign right, we'll know where he saw our ad. Was it the outside billboard ad in Earl's Court, or did he respond to the TV ad we ran on the tennis broadcast, or the ad in the Sunday newspaper. Now we get digital footprints about what the interests are.
In Helsinki Finland over half of all public transportation tickets are paid by mobile. In Japan 44% of Japanese consumers click on ads on mobile. In Kenya one in five bank accounts is a mobile account. In South Korea 45% of all music is sold directly to mobile phones. As the phone takes on more functions and gives us more services, we also leave ever more digital footprints.
These are actionable. Our digital identity is not actionable. But digital footprints are very actionable. Also I want to thank our friend Tony Fish (co-author of Mobile Web 2.0) who opened up this line of thinking last September and spoke of it at the Mobile Web 2.0 conference last year.
If the digital footprint is exciting and valuable and promising, it is still old-fashioned "1.0" thinking. Now lets move to "2.0" thinking. From customer insights 1.0 to customer insights 2.0. Now adding the social context of our behaviour.
By social context, I now move beyond what this one person is doing, and consider rather who is involved when that is being done. Not knowledge about the Who or the What (or when or where), this time we want to know "with whom", Lets go through some examples.
We observe the behaviour of the phone. We have the daily movements. We notice that there is another phone which moves the same way, starts off the same place and moves to the same destination every day, returning at the same time. We track the two phones, they are not using identical patterns - could be two phones by the same user - remember 28% of all phone owners on the planet have two or more subscriptions. But this is not a case of that. In our tracking we find that the second phone is also making movement patterns separate from the primary one we monitor.
And we notice the two phones call and send messages to each other. We wouldn't be doing that with our own phones.
So its two separate people. Lets track them more. It could be that they travel part of the way from home together, then they split, and at the end of the day they come together and go home together. This could be a parent taking the child to school or daycare, and picking up again at the end of the day.
Or it could be two phones travelling together very regularly coming and going on weekdays but a lot of separte movements on weekends. Two work buddies, who live near each other who commute together (or might be romantically involved).
Or it could be two phones who really seem to live and sleep together - husband and wife working at same employer.
Etc When we add social context, we get far more valuable information.
Now we know MUCH more about our target. If we interviewed him or her, and this is for example an adult, who has gone back to school because he was unemployed, the person might be ashamed to admit he's gone back to school. Or it might be a teenager who "should be in school" - and whose parents even dont' know that he stopped going to school months ago and has taken a low-paying job, etc. The digital footprint won't lie. It is far more accurate than anything else.
Then lets take a passion. Lets say our target did not vote for Big Brother on TV. But we know the target is always home when Big Brother is on (is not making the pattern of movement to the nearby pub, the person does twice per week, and is not shopping for groceries etc). We don't know what channel he is watching. But we use social context.
During Big Brother the person sends more text messages than normal. And these text messages go to a clear pattern of five other people. And two of those five voted in Big Brother. And when we map out the total network of those five people during Big Brother broadcasts, we find them interacting with 15 other friends, giving us a circle of 20 friends who all do a lot of messaging during Big Brother, and out of those, 9 have voted for Big Brother.
We can be sure, that our initial person IS a fan of Big Brother, and is religiously watching that show, even though he (or she) has not voted in it. Because of social context. Those other friends are also all Big Brother fan. All 20 prioritize their lives to be able to watch that TV show, and half of them vote in the show. Our target communicates with these "clearly Big Brother fans" regularly DURING the show. Only another fan of the show would follow this pattern.
We get so much more information out of social context. Take me for example. I have a totally consistent, regular-as-clockwork peak of alternate Sunday messaging. For 9 months of the year, for over 10 years already, every other Sunday, I send a large batch of SMS text messages. When you map that pattern to the local times, its clear to see, it fits perfectly with the Grand Prix (Formula One) racing. If the race is in Bahrain, four hours before UK Greenwich Mean Time, then my peak will be from 9 AM to 11 AM London time; if the race is in Tokyo, it will be even earlier, and if the race is in Rio de Janeiro, my messaging peak happens in the evening, etc. But clear pattern.
Now, if my mobile operator wanted to find Formula One fans, they could use that pattern and detect who fits that pattern - and then again, social context - who they message with. Now, take the total group that emerges - and map it against these phones turning up at any Grand Prix race, whether Monaco in May or Silverstone/UK in July or Spa/Belgium in September - if a few of this "circle" of phones travel physically to a Grand Prix race, then certainly all within that group of contacts, who send messages during the races are real F1 fans (it could be the opposite - say the wives of F1 fans, who think this is the safe period when their husbands are glued to the set, but then my guess is the social pattern would be more direct phone calls than just messages, ha-ha).
Social context. Give me two months of your phone behaviour, and the total traffic in the network, and I'll know more about you than any marketing manager could ever capitalize upon.
And again, we have not even gotten to the new services and applications enabled on the phone.
Now lets bring social context to our consumption. This phone number 0123456789 is at a shopping district - against the normal pattern of going home from work. Then the person accesses Google or EBay on the phone and looks for a plasma screen TV price. We can be pretty sure the person is at an electronics store pondering a purchase of a new plasma screen TV. Then the person calls a number, speaks for a while, and stops shopping, goes home.
We track this pattern for a while, and we notice that our mystery phone gets lots of these calls during the month, from random callers in the network, and often after those calls, the caller stops circling in the shopping centre and heads on home. We have found a digital gadget guru. This is the guy who advises everybody about the best electronics to buy. Social context.
Then take our adidas example. Our target indicated an interest in adidias. But what kind of interest. We track his calling and messaging patterns, and find that there is a group which has several members (but not our target) who regularly go on clearly mountain-hiking patterns (fast movement to remote hilly districs ie movement by car, then very slow movement in the mountains for a couple of hours, then return home again fast; hikers). The guy has lots of friends who are hikers. Now we map the parallel behaviour, and we notice that two of this group, tend to go slow-speed moving (ie hiking/walking) in the city on given Saturday and Sunday mornings. The guy likes to go powerwalking or nordic walking or whatever, with his friends. Social context again.
I've often used the example of Madonna's world tour. Imagine Madonna at the new Wembley Stadium. When she performs there, almost all phones present at her concert are die-hard Madonna fans. The value of knowing their phone numbers, is enormous. These people stood in line to pay top prices for seeing Madonna live, these are MORE intense fans, than just those who buy her CD's. Now, if we know the social context of why these people gathered this Tuesday at Wembley Stadium (when on Thursday it might be a football game and on Saturday it might be a rap concert etc) then we have actionable data that is far superior to any market research data.
And then - imagine the power of viral in this context. If we know the person's interests - whether its Formula One or Madonna or Nordic Walking or whatever - and we give a viral bit of marketing, an advergame for example - to these hard core fans. They will distribute it, and if we track who they give it to, we'll gather the full social circle of who is involved and passionate about that given interest.
This is the future of marketing intelligence. Its no longer demographics. Identity is not worth collecting. Lets safely secure that with our customers, promise them we won't mine their identity. But the digital footprint, that is valuable. And the social context - Like Alan Moore says, this is the Black Gold of the 21st Century, the biggest prize. We can only discover social context accurately via the mobile phone, but the companies that build upon this dimension, those companies will seem like "reading our minds" in how accurately, cannily, they will serve ever better services and products and offers and campaigns for us.
And a reminder. I wrote my latest Thought Piece on mobile advertising. It briefly also touches on these issues. If you'd like a copy of the Thought Piece, please send me an email to tomi at tomiahonen dot com and I'll send you the pdf document by return email.