The Economist had a great special report last week about data and it had me thinking. Yes the world is changing enormously as all sorts of companies compile ever more accurate data on us as consumers (and as patients, as voters as students etc). The companies (and government, schools, hospitals) can potentially deliver better services and products for us with this data. And they are increasingly collecting it through automated means, often via the web.
THE DATA COLLECTION IS PRETTY PATHETIC ACTUALLY
The Economist article had great examples of how the various industries today are taking advantage of vast data amounts being collected on their users. And it was all good stuff, I strongly recommend reading the article. But I was left thinking, that while the Economist did often mention mobile, it did miss the big story. It is a bit like going to the local high scool winter sports contest, and reporting on that level of ice hockey as if it was the world's best, while there are Winter Olympics going on, and broadcast on TV at the same time...
So, first to be clear, it is a very good thing that companies and other organizations like governments, hospitals, schools etc, do collect data and then mine them to get 'business intelligence' to better understand their customers. That is all good and prudent. And we can have truly magnificent 'improved' services for us, based on that data being collected, as I showed last week with the commentary of Amazon, and how its recommendation engine has made advertising seem like valuable content to us.
But if individual companies or organizations try to collect data on us, it will be a poor picture of very modest value and usability. Like many examples in the Economist article, saying how a museum had been able to target its advertising more accurately based on 'demographic' data. This is better than nothing, just like high school level ice hockey - under coaching and with facilites and organized competition - is better than totally volunteer street level hockey. But it is nowhere near Olympic level hockey.
So now lets compare. Lets say we take a chip and a memory device and perhaps a connection, and build an automated data collection system. Lets say we want to track usage info on the car. So we have a family with a BMW and we then see that the car is driven most days on pretty much the same route in the morning and evening. In mornings its via a detour to the employment place of the husband (the detour means he drops off the kids at school) and some days it comes back the same way via the school, but on other days comes home directly, and occasionally makes other routes to get back home (ie goes shopping etc). Some days the car does totally different patterns but we have no way of knowing who drove it. We get to know how fast the car is driven, how it is accelerated and how the breaks are applied, how much fuel is kept in the tank (always nearly full or often nearly empty) and how often it is driven etc. Very much valuable data, potentially. And its all good. And there are tons of 'experts' in data mining who can think of clever uses for this data.
Now. So far so good. But do we know who is there with us? No. We may make a guess that its the man in the family, the father, who bought the car, and perhaps we can find out that he is employed at the nearby factory as a manager and we assume its the man driving the car daily (but could be that both he and his wirfe work at the factory, and he's not even IN the car, that he drives a Jeep for example, that he bought this BMW for his wife...) What of the car used in the evening on a Friday. Do we know if it was the father or the mother or perhaps driving-age eldest child who used the car? If we have some sensors on how the car is driven, it won't capture who is driving it. Perhaps we can have electric seats with family settings and find that the shorter person drives is occasionally and the taller person drives it most of the time.. Of do we want our cars to ask us to 'identify' ourselves before we get to move it? Personalized car keys - messed up the moment the husband says 'honey, I can't find my keys' and the loving wife gives him her keys.
Now compare to the phone. We don't share our phones, but we take them literally everywhere. If we knew the phone numbers of each family member (and for American readers, yes its normal for 7-year-old first-grader kids to have their own cellphones, you will shortly see that in the USA as well) -we'd have perfect knowledge of the drivers and passengers (if any) and the driving patterns. The cellular network can collect this level of data automatically with no further data from us. We don't need to call anyone, we don't need to send messages and we don't need to ask for GPS location data. We get at least 'city-block' level accuracy of all mobile phones turned on in any cellular network, just from the network data (if the operator/carrier decides to track that data and collect it - they all do it for calls and messages, they can very easily do it for the phone's behavior as well)
The owner of the phone - BMW buyer, the husband in this family - has his phone registered with BMW. We know if it is the father who drives every morning to the office or perhaps its the wife every morning (daddy might have a Harley haha). We'd know if its one or two kids in the car with daddy (or mommy) every day. We'd know if both parents are in the car. We'd know which person was driving the car in the evening on Friday.
First, from the pattern of daddy's phone, we'd find out who lives with him. It does not matter if the adult woman who also drives the car occasionally - is legally married to him or perhaps just a long-term live-in girl-friend (or even if its two guys and thus its his 'partner') . If we monitor those who use that car, and their sleeping patterns for a month, we know exactly if the 'partner' lives with the man, or if the partner is only occasionally visiting the home and lives in another place. We'll know how many phone-owning kids live in the household and/or any grandparent age people in the household (with phones).
Next, we can easily follow all those and see how they move, to determine if they own a car (or perhaps, a motorcycle, wtih a little bit of extra analysis, determining how the vehicle moves in traffic, to determine whether its a car or motorbike)
And back to our BMW owner. Compared to the BMW own analytics built-into the car, with mobile, we'd still know how fast the car was driven - but very relevantly, we'd see when there is a single phone of a driving-age person in the car which family member has which driving pattern. Who drives as fast as the traffic - stays in the lane - or who is the crazy driver who is constantly overtaking others in traffic and potentially causing dangerous situations. (from the network we would get every car driver's phone and its movement speed - its very easy to spot which phone is clearly moving on the same route but faster than the traffic - ie either on the bus lane - illegal - or weaving in traffic - very dangerous).
Now lets take it a bit further. What of the passenger's seat. Unless we install a 'sensor' to every seat, we won't know how many people are in the car at any time. But with phones we know. Lets say the husband is unfaithful - has a mistress. So he drives his wife to her work, then goes to his office - but some lunches, he goes out secretly and meets with his girlfriend. And every so often SHE is sitting in the car. While in any surveys of how many people are in the car, the husband would never admit to having a secret lover, the phone behavior would prove this conclusively. Is the passenger's seat used? How important is say extra air bags to the passenger seat in the next purchase. If the father lives alone and has no girlfriend, odds are he won't want to spend more money on extra airbags to the passenger seat. If there is an important woman in his life, he probably would. What should BMW try to sell him? If we know how the car is used, how many people in it, how it is driven, etc, that is actionable info, not how old I am or whether I'm married or what postal code my house is in etc.. Demographics become quaint at this point, when we have real behavior-based and social context-aware data.
Note this is all info the network already collects today, automatically, every network, everywhere. The networks do not yet offer this info in any way, it is not monetized. But think of BMW, how much more info they'd get if they were to catalog the phone numbers of their owners, and then be given permission to track those phones by their movement patterns.
TOMI THATS INVASION OF PRIVACY
Sound like invasion of privacy? Sure it does. But first - remember the cookies on your PC. Nobody ever asked us to approve these spy-bots that we all now are infested with. Secondly, technically, every mobile phone network has had this capability for almost two decades already. Thirdly, they already do use this - with 'lawful intercept' ie when fulfilling a given country's legal system, the police request the data and the operators/carriers provide it. Usually this requires a 'court order' type of legal document for the police to gain this access. But the 'snooping' is totally technically valid today and done around the world for such uses.
But lets think Amazon - we 'love' the recommendations that Amazon makes for us. Why could not BMW be able to provide better services for us in exchange for this user info? Jonathan MacDonald teaches us that we need to be open about our intentions. But it could turn the whole marketing budget equation upside-down. How much do car makers waste money in advertising every year. BIllions of dollars. Most of it, by far most of it, goes to serving ads to people who would never be buying their car. Most will not buy a new car, period. of those who would, most would not consider that brand. Of those who would consider, often the ad itself is unsuited for that given person. Why do I see ads for Honda or Volkswagen or Ford or Cadillac? I am an Audi guy. I might perhaps perhaps perhaps might consider a BMW or Mercedes or Porsche or Saab. I would never ever in a million years buy a Honda or VW or Ford or Cadillac. If the car makers could know what kind of car (and in my case, Audi - would be a Quattro four-wheel drive model, only) a given customer is interested in, they'd save tons and tons and tons on their ad budget. They could throw 90% of it away at least.
Nwo lets go back to my BMW example. We want him to give permission to track his phone. Why not offer a 'roadside assistance' VIP package to buyers of new cars, that is based not on some OnStar type of service, but as a mobile phone data service. Imagine how much MORE powerful the data now becomes. I not only track how 'my' BMW vehicle is moved - and can quickly map out his family etc - btu get this - I will see automatically when my 'BMW owner' is moving 'at car speeds' - in another car! Imagine if American Express could get this data about Visa and Mastercard. When is an Amex card holder using a rival card? But yes, with mobile, we'd easily get this. We do the plug-in hands-free package (free) in the car, so the BMW owner will always plug his phone to his car and re-charge the battery and use it when driving the BMW. And then we monitor both the car connection and the registered phone. If the car moves but has another phone 'driving' it, we'll know. And again, we can track from the network if this is the same person with two phones - then those 2 phones go everywhere together all day - or if its 2 people - the phones split apart during the day when the wife and husband separate at some points..
Is there that 'other car' in this same household? (could be the wife's car) or perhaps its just a friendly neighbor who gives a ride sometimes and the two neigbors 'car pool'. If we know the phone number of the BMW owner - we can within a month's of data of phone behavior clearly establish who are the other phones who live in the same household (how many kids and is there a wife also or perhaps the husband has divorced her wife and lives alone with the kids).
Ok, now lets say there is another car. Wouldn't it be awesome to know WHAT BRAND the other car is? Again, not absolutely certain, but if that other car goes twice per year to a registered dealership of say Audi, we know the wife has an Audi while the husband drives the Beemer.. Can you imagine how much more valuable this type of info is that 'what age is he' or 'she' or 'is he married' etc. Demographics become totally useless as data, meaningless as being so inaccurate, when we get mobile phone based behavior data.
WE ARE NOT THERE YET
We are only starting on this journey. But regular readers of our blog know that Alan's second book Social Media Marketing (See here in the margin to the right of this blog) deals with this area. Our friend Tony Fish has written the definitive book on My Digital Footprint which outlines the differences between the meaningless digital identity and the incredibly valuable digital footprint.
Most mobile operators are still clueless as to how powerful this data is. They are not even capable of segmenting their OWN customers to this level of detail, far less to sell you any data on them. But advanced operators from Scandinavia to Japan and Singapore are already on this path, far far along the way. And don't think 'privacy' is any protection. We lost the battle with cookies years ago. The mobile operators, and various industries (like car makers in this example) will get our permission - easily.
Yes, it has to be permission-based if you want to market directly to me - totally yes. And every brand should NOW be seeking that permission, when the consumers are still pretty ignorant and don't understand the full value of the digital footprint. (and when we are not bombarded by many mobile ads every day) But yes, even if the consumer did not give permission, the operators can totally legimately sell the data if it is 'aggregated' ie that our names are not included. But take the above example. BMW would probably be quite interested to know how many of those BMW owners, who own 2 cars, have an Audi as the other car (as would Audi). Note that as long as the mobile operator does not reveal my name in this example, they are fully allowed to sell BMW the 'statistical data' that tells that for example 7% of BMW owners have an Audi in the household but only 2% have a Jaguar (or whatever the data may be). This can be easily done to city and township level with no infringment to any privacy laws...
And more than that, to know of the family with both an Audi and BMW, which car is used in the weekends when the husband and wife go out together, etc... Imagine the power of that info! If you are Audi and you know that of families with both an Audi and BMW, 76% use the BMW as their preferred vehicle. Wow, that would be valuable info. No sensor in the car you sold, will ever tell you that, but mobile networks will collect that data easily, already today... With just a little bit of analytics and logic and consumer insights, this is child's play to generate and then to monetize. Like Alan Moore says, this is why mobile phone user data is the 'new black gold of the 21st century'.
NOW THINK FUTURE OF MARKETING
Now, lets think the future. What if the car makers all know this and did this? And then they'd build clever marketing plans around it. So, for example - BMW fanatic fan number one - we get the owner of the BMW who loves the Beemer so much, he decides to become an advocate. He signs up to BMW's enthusaists mobile service and gets tools to spread the word. One is to send invitations for other phone owners to sign up to the BMW fan club. The new fan gets some 'BMW loyalty points' for signing up (and then further points potentially for staying loyal - which eventually of course would translate into modest discounts to the next purchase of a BMW). But our fanatic - he'd also get points for every fan who he has invited to join - who then joins and remains in the club for say several months...
Then there is the 'design the ultimate BMW' contest - have the fans each design a virtual car of their ultimate BMW. It could have two classes, one in his own model range, and another of any BMW platform car. Limit the designers to those who own a registered BMW car, so we won't get non-owners messing up BMW's model lineup haha. Then, if you owned a 3 series, you could design the ultimate 3 Series (by any BMW official accessories and options) and make another entry in what for you would be the ultimate car if price was no object, say the 7 series for example. Make this a contest, have any registered BMW fans (ven those who don't own a car) be allowed to rate the cars, for each category give out the year's best car design to the designer.. Saves tons of money on market research because obviously the top rated car models would be also shipped to the dealerships in their model ranges.. It could even have a custom name after the designer. Imagine that, my car would ship in thousands of units globally, each having on its model description in the back of the car "BMW 3 Series - T Ahonen Special Edition" or something like that. And imagine how passionately those who would receive this honor, would then become in advocating their model to anyone haha..
Now - how valuable is this data to BMW. Tens of thousands, soon hundreds of thousands, then millions of fans would design their Ultimate Driving Machine. When this registered BMW owner next comes to test-drive the new model, the dealership will know if he prefers the comfortable seats or the sports driving package etc. And best of all - at some point, the dealer will have a model in his shop which will be very close to - perhaps even identical - to the 'dream model' that this fanatic has desired (remember for one winning model design, there are hundreds of thousands of BMW owners who designed their own dream cars that did not win). Why not send the surprise MMS message to that BMW owner, with the picture - saying, "your dream BMW exists today, would you like to come and test-drive it - and keep your dream car for a day?" How likely is it that this BMW owner will buy THAT very car? Especially if the dealer makes a special offer of it, taking into account the current car, its depreciation etc...
TEST DRIVE MY CAR
Then, another class of fanatic - the 'test drive' friend. Why not award loyalty points every time to one owner of a BMW, when he or she 'loans' the car to any friend. The only requirements are, that the friend then register their phone with BMW's fan club - at least for the duration of the test drive - and no owner gets more points for lending the car to the same person again. But yes, why do this? Number 1, BMW knows people like the way the car drives. Its likely if the 'fanatical' BMW owner picks candidates, these will appreciate the car. Secondly, BMW gathers phone numbers - and permissions to market to - prospective future buyers. Now, the person who offered the test drive, gets loyalty points if the candidate registered with BMW. Then the person who signed up, will get loyalty points for signing up. The test driver will get more points for driving that car (its possible he/she was already a member of the club). Then if the person who did the test drive comes to BMW to try any BMW car, he/she gets more points for test driving and obviously these points will turn into a discount on the final car. Meanwhile - our Fanatical BMW owner - he/she gets points for every person that tests his/her car ( and signs up to the BMW fan club) - but - he/she gets bonus points for any friend who then buys a BMW.
Third - remember we register the friend's phone - and we have the 'loaner' tell us how long the car was given to the test drive - so if the friend loans it for a weekend or an hour or 5 minutes, this is valuable info, why? Because we can go back in the phone data and see how the new registered phone moved for that day or hour or 5 minutes - we get a pattern of how the 'friend' was driving that BMW. It gives us good clues as to what kind of driver he/she is, and perhaps what kind of features the person will appreciate in the car they will eventually drive in our showroom! Awesome...
THEN YES, MONEY TO THE FANATIC
I would think that this should work out rather well in the 1%-5% rangesin loyalty points when converting to a cash discount on the next new BMW. The points would be actual 'BMW cash' redeemable at the BMW dealer - after the final price is set, and the points would come straight from BMW headquarters (out of their global advertising/marketing budget). Meaning that in a perfect case, if you were truly a great 'salesperson' and did all the steps, and you personally were credited with the sales of 20 of your friends to buy a new BMW - you'd get a free car yourself !
Sound crazy? No, this is just the recommendation model we already have with for example Amazon partners. The main problem is the collection of points and correct attribution of the rewards, but something like this will come to car sales sooner or later, and the model will be built around our mobile phones.
Think about it - if there really are 'fanatical' BMW fans who actually 'sell' the car to their relatives and friends - then shouldn't we correctly credit them with that work. Reward them? And there need be no total limit to how big a benefit they can get out of it. If they do the real sales work, that means less advertising and less sales effort at the dealer, easier money for BMW - of course some kind of 'commission' should go to the guy or gal who did the recommending. The bigger the reward potential would be, the more our fanatic would work to be our salesman/woman. Most would not sell 10 cars, but one might sell 20 or 30 or 50.. Then that person deserves a new BMW every six months haha, why not? He/she is doing real sales on our behalf, except that far more accurately, he/she knows his/her friends, who will really want the car, appreciate the car etc. (plus we dont' need to pay that fanatic a monthly base salary as we would our regular dealership salespeople, these would be on pure commission-basis sales haha).
But think - honestly - we've all been in that situation at one point or another, where we had a friend considering a car, and we advised to try a car we liked (and thought our friend or relative would also like). Or if not car, then mobile phone or personal computer or TV set or brand of eye shadow haha, whatever it is..
And it doesn't end there. How many times have I been recommending British Airways to my friends. I love the airline. I've been collecting friends to sign up to their frequent flier program etc. But do they know that, no. They do treat me well - by my actual flying pattern (because I am on their frequent flier plan and fly enough on them) but they have no idea how much more I am "selling" them and their services when I talk about them with fellow travellers from time to time. Now, if BA was really clever, they'd be sending me occasional 'gift bonus discount' coupons via mobile (as MMS haha) that any BA traveller could use to redeem. So these would be discount coupons not for Tomi Ahonen's use but that any other traveller could use. 5% off any flight.. Why not.. Send a new coupon valid for the next 3 months, say once every month. Not too many, but a few every so often. Any coupon could only be redeemed by one recepient once, so I could not set up a pattern where I send my monthly coupon to Alan and he sends his monthly coupon to me haha.. Any one person could only use one of these 5% discounts per year, or something like that..
Now, how often do I know of a friend heading here to Hong Kong (often from or via Britain). I'd be happy to send some BA discount coupons to them. Even if only a small part of those coupons were ever used, it would be often a 'change' in the flier's airline pattern - where they'd be sampling British Airways. Now, if BA was really really clever - they'd then track my coupons and the passengers who took them - and where they find the phone number is a new first-time BA flier, (ie not already an existing BA frequent flier haha) they could offer these passengers (on a one-time basis, as they knew the phone number) an upgrade on one leg of the journey, or if the class is full, offer a one-time invitation to the lounge, etc. Why not - this is a chance to capture customers away from rival airlines, in a totally new way. Because BA get the phone number, they can limit this and not give the trial more than once, but as BA's offer is not the discount brand, but a luxury premium brand, then they should try to get as many prospective travellers to trial their lounges and premium travel cabins as possible. So next time that customer will know to consider theri service haha.. And of course at the point of redeeming the coupon - make sure the person signs up to BA's frequent flier club with some gift bonus miles to start them off haha... (this would also reveal if that passenger already is a FF passenger but gave another phone number to get this offer..)
Meanwhile as gradually some of those forwarded discount coupons are redeemed by my friends, BA could reward me for doing their PR - easiest by giving me air miles, say a quarter of the number of air miles that my friend bought with my discount coupon. If it was a short European flight in economy class, not a big thing, I 'earn' a couple of hundred miles. But if it was business class to Asia, this is BA's profit-making engine, business class - then yes, of course, very nice bonus thanks to Mr Ahonen, I earn several thousand miles as a thank-you. Plus for time there was a redeemed coupon that I had sent, I would get 2 bonus coupons to send to other friends, above and beyond my normal allotment..
And I'd love to do it (...with any airline haha, hint-hint..)
This is again obvious what will happen in the long run as companies learn that mobile phones are the optimal engagement marketing platform, and they seek to develop good engagement marketing campaigns. And as they start to gather mobile data, and as the 'data mining experts' (haha, sorry, I have to laugh when they still mention demograpics in the 2010s decade, its that old and obsolete thinking..) start to get to grips with mobile phone based behavior data.
A side note, did you pay attention that the data mining giant Sybase had just deployed a mobile solution with SAP the giant of Customer Relationship Management software (Sybase is also very big in mobile, messaging, SMS, MMS, mobile commerce etc, via their mobile division Sybase365). It was just announced this week. Some companies see the value of where this is going and can be very big in this space soon. If I was an airline or bank or retailer or hospital - or BMW or Audi haha - I'd make sure I'd call these guys - Sybase and SAP - now. They really do 'get it' haha, in this space and opportunity. Not like so many others mentioned in that Economist article, who still think 'demographic' data is the holy grail haha.
I have added a blog discussing the deeper insights, there are actually 3 levels of consumer insights, that only mobile can give us today. I show them in this blog 'For Your Eyes Only'