20
www.ctrl-shift.co.uk July 2012 Summary 2 1. Introduction 3 2. The study 4 3. Background 5 4. Our approach 7 5. Our findings 9 5.1 Inferred interest 5.2 Loyalty, retailers and deduced interests 5.3 Get customers to volunteer interests 6. Summary results 17 7. Conclusions 18 8. Comment 19 Don’t target me - ask me (and make it fun) a Ctrl-Shift report

"Dont target me ask me" ctrl shift report

Embed Size (px)

DESCRIPTION

 

Citation preview

Page 1: "Dont target me ask me" ctrl shift report

www.ctrl-shift.co.uk

July 2012

Summary 2

1. Introduction 3

2. The study 4

3. Background 5

4. Our approach 7

5. Our findings 9

5.1 Inferred interest

5.2 Loyalty, retailers and deduced interests

5.3 Get customers to volunteer interests

6. Summary results 17

7. Conclusions 18

8. Comment 19

Don’t target me - ask me(and make it fun)a Ctrl-Shift report

Page 2: "Dont target me ask me" ctrl shift report

Findings

• Consumers are excited at being able to unclutter theirlives of irrelevant messages.

• Consumers are realistic about their privacy and willreveal information if control is fun and not a chore.

• Instant gratification is key, be that through gamificationmechanics and/or something immediate in return fortheir efforts.

• Inferring a consumers interests based on their digitalhistory is a thing of the past, only with the consumerinput can marketers be forward looking.

Our study

• This study looks at how giving control of “what I aminterested in” to consumers delivers a betterexperience.

• The study, based on ten in-depth assessments of eightservices, looked at three ways of serving content(adverts, recommendations and offers).

• It assessed three ways of the supplier evaluating whatconsumers were interested in.

• First, inference. Based on browsing history, the servicesassessed - Facebook, Google and Bluekai - all deliveradverts that are targeted against those interests

• Second, deduction. Based on shopping history, theservices assessed – Amazon and two loyalty schemes –offer recommendations or discount vouchers relevant topast purchases

• Finally, volunteered. These new services – VisualDNA,Dealboard and Hunch – all turn the problem around.They ask consumers what they are interested in andthen deliver content against that interest.

• This third group do this not by asking for traditionaldemographics but by making the exercise fun – and bynot asking for personal information such as age, addressand so on. Instead they use non-personal information todeduce this. For example, your personality, the films youliked or brands you admire.

• Ctrl-Shift used a methodology that it has developed forthe UK government’s Midata programme – TACT:transparency, access, control and trust.

• The three services that used volunteered information didbest in our tests (although Amazon gave them a run fortheir money).

• Surprisingly, perhaps, the loyalty schemes didn’t do well– “they are clueless about me” said one assessor.

• And although the web-browsing trio did worst, there wasa genuine acceptance that, while information remainedanonymous, this profiling was acceptable.

• However, with advertising on Facebook et all, clearlycausing huge amounts of wastage (9,995 adverts are notclicked on), there is clearly a problem to solve.

• Using the information that people are willing tovolunteer could solve this problem – delivering contentto people who express their interest is marketingnirvana.

• But perhaps as importantly, this study revealed thatdoing this in fun and engaging ways can provide a newservice – analytics for consumers.

• Allowing consumers to work out why they buy, use orlike products, really would be a control shift – a gamechanger indeed.

Summary

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

PAGE 2 © Ctrl-Shift 2012

Page 3: "Dont target me ask me" ctrl shift report

www.ctrl-shift.co.uk

PAGE 3 © Ctrl-Shift 2012

Information and communication technology is evolving rapidly. It is common for virtuallyeveryone to carry a mobile phone and they are quickly acquiring smartphones. Thistransition is enabling a step change, or in academic terms, a paradigm shift in the evolving relationship between technology, enterprise and the individual.

Individuals are no longer limited to one-to-one voice ortext communications. They broadcast views, photos andjust about anything else you could imagine through socialmedia, shop without having to enter a shop, search forjust about anything and find it on a map and then getdirections to it through their GPS, all at the touch of abutton – or should we say ‘the touch of a touchscreen’ asthere seems no end to the trend towards hand-helddevices.

However, this also means there is a lot more informationabout individuals available for exploitation. The newtechnologies can glean our transactions, interactions, eventhe language we use, and attempt to infer or deduce ourfuture preferences.

This data collection capability has not gone unnoticed by the commercial world that is in the process trying tofind ways best to use this resource. Amazon uses pastpurchase behaviour to suggest other items. iTunes doesmuch the same.

For quite selfish reasons, firms are finding that if they letpeople volunteer what they want and control theinformation companies receive, they benefit. It enablesthem to target precisely the right people with the rightproduct at the right time. Known as Volunteered PersonalInformation (VPI), it is a big improvement over thededuction and inference-based market research that forcesuppliers to employ mass market techniques that hit boththe interested as well as the increasingly annoyed.

However, what about the individual? How do they benefit?How? Or are they just being manipulated?

It is the individual’s experience of this emergingrelationship with businesses that we explore here. Is therea better way of serving content to people without themfeeling that they are being spied upon, by using marketersbest guesses, by using complex mathematical algorithmsor simply by deluging the customer with special deals?

1. Introduction

Don’t target me - ask me (and make it fun)

Terms like exploit highlight a core challenge in thismarket. Businesses see data as something that theycan use to their gain. Period. As the World EconomicForum called it “personal data is the new oil’. Wrong.Businesses that exploit their customers don’t engendertrust or loyalty.

Page 4: "Dont target me ask me" ctrl shift report

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

PAGE 4 © Ctrl-Shift 2012

We see three macro trends at play in this market:

1. The rise in the volumes of data available (so called Big Data).

2. The lack of efficiency in the advertising model.

3. How technology is now literally in the hands of the customer.

Each is a driver of change but combined they create amassive opportunity for a re-alignment between therelationship between buyers and sellers.

In order for VPI to work, people have to take time tocontribute information. If there is no benefit in eithersaving time or there is a financial incentive, they simplywon’t bother.

This in turn has driven suppliers to offer just suchincentives, either saving them time searching for goods,they are presented to the willing consumer often with aspecial voucher or discount.

The wonder of technology’s evolution in a market economyis that it is driven by the needs and wants of consumers.Applications that appeal to consumers win out over thosethat cause despair. The soulless face of technology meetsthe cruel realities of the market and the consumer seemsto be benefitting.

We wanted to find out for sure and better understand howthe market is responding to this new phenomenon.

We focused on five questions:

• Do consumers know what information is held about themand how it is used?

• Are they concerned by the information held about them?

• Do they like what suppliers infer or deduce from it?

• Are they willing to let suppliers know what they arelooking for more directly?

• How might consumers benefit from the relationship?

2. The study – Emerging VPI systems and their effect on the individual

Page 5: "Dont target me ask me" ctrl shift report

PAGE 5 © Ctrl-Shift 2012

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

Citizens in this consumer-driven world searchfor information on their interests and areinspired, some might say, bombarded bymessages and offers. So many in fact, they arechallenged, often irritated, to have to sortthrough the chaff to find the wheat.

Reason for the deluge?

For the last 20 years (and more) companies have beenlooking to collect more and more data about theircustomers. The more you know about your customers themore successful you will be. However responding to aperson when he or she doesn’t want what you have tooffer, creates waste for both producer and buyer.

Unable to target precisely, suppliers instead used ashotgun approach that hits far more people that could useor want the approach simply because supplier didn’t knowthat much about them.

Technology – part of the solution or adding to the problem?

New media, customer loyalty schemes, online commerce,location-based technology and social media have openednew channels both to collect data and to send messages.

It has enabled suppliers to make ever more preciseinferences and deductions about consumers. Whereasbefore they might only know their gender and age or whatthey could ask about them, suppliers can now gain insightsfrom their behaviour.

This has still lead to discontent on both the parts withsuppliers still unable to know sufficiently well when tocontact people, and citizens approached with the wrongmessage or at the wrong time.

A New Approach – Volunteered Personal Information (VPI)

Not only are people switching off from mailing lists butthey are also finding ways in which they can take controlof the relationship with their suppliers. Telling them whatthey want or think - think Amazon reviews or TripAdvisor.

As information starting from customers has exploded, sohave the business models designed to take advantage ofthem. A new paradigm is emerging mainly due itsattractiveness to both suppliers and consumers –Volunteered Personal Information. In other words, there isa third channel opening up between suppliers andconsumers, and this time customer-initiated.

Rather than businesses trying to collect as much data aspossible, there is an alternative – ask your customers whatthey want, right now and in the future. Instead of guessingthe sorts of things that people are interested in, why notfind quick and easy ways of getting them the offers they want?

3. Background

Half the money I spend onadvertising is wasted; the trouble is I don't know which half.John Wanamaker.

Action Conclusion

Inferred You’ve searched We should advertisefor babies toys baby equipment

Deduced You’ve bought nappies You must have a six for the last six months, month old childalways increasing in size

Action Conclusion

Inferred You’ve searched We should advertisefor babies toys baby equipment

Deduced You’ve bought nappies You must have a six for the last six months, month old childalways increasing in size

Volunteered I’d like to buy a pushchair Let’s send some offersfor a 1-year-old

Page 6: "Dont target me ask me" ctrl shift report

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

PAGE 6 © Ctrl-Shift 2012

And governments across the world are supporting this.From the US’s Smart Disclosure, the UK’s midata initiative,the Australian health records initiative to the WorldEconomic Forum’s Reinventing Personal Data project, thegovernment agenda is to help businesses create economicgrowth by helping customers make better decisions.

This new era, the control shift, is already beginning todevelop its own lexicon. CRM could become VendorRelationship Management, which emerged from Harvard;the principal academic in this area, Doc Searls launched inApril 2012 a new book describing a new economic age –the intention economy; and so on. And we at Ctrl-Shifthave added a four-letter acronym as part of our worksupporting the UK government’s midata programme – TACT.

Transparency, Access, Control and Transfer providesbusinesses with four pointers to help them benchmarkthemselves. First, make it clear what is happening to datasupplied; second, give people access to this data; third,give customers control over that data and finally, if theywant to leave, find a way that they can transfer theirrecords out of your system.

3. Background CONTINUED >

In a world where people are takingback ownership of their data, whatare the threats and opportunities for our industry and us.Global supplier at a workshop in June 2012

““

Page 7: "Dont target me ask me" ctrl shift report

PAGE 7 © Ctrl-Shift 2012

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

Ctrl-Shift has:

• Reviewed the existing literature andinformation on the topic;

• Conducted panel sessions with a cross-sectionof the population to discuss citizen-initiatedapplications and the information already heldabout them and had panellists rate severalapplications side-by-side;

• Assessed the experience of a cross-sectionparticipants; and

• Interviewed in-depth a cross-section ofparticipants.

The Panel

Approximately 20 people were asked to take part in thestudy. Each was selected if they had a Facebook page, useGoogle search, had an Amazon account, some sort ofloyalty scheme and used an iPhone. For various reasons,people dropped out (eg “I didn’t realise I needed aFacebook profile” or “I gave up when there was no Googledata on me”) as well as the researcher’s curse – peoplewho promised to attend and didn’t.

In the end, a self-selected panel of ten people emerged.They ranged in age from 18 to 57 (average 36), and were amix of males and females, married and single, withchildren and without, PC users and Mac users. They had awide-range of views on privacy coming into the projectand also participation in loyalty schemes.

They were asked to assess a number of services on a rangeof criteria. The panel was asked to look at these servicesin the same order (because we wanted to take them on ajourney from the familiar to the unknown). They wentfrom Facebook, to Google (and online search) to loyaltyand finally to some new services. The core continuumthrough this journey was the profile that the services hadbuilt up about the user.

For this project we have adapted TACT (Transparency,Access, Control and Transfer), using this same approach tolook at the spread of services covered in this research. Wehave, however, replaced Transfer with Trust as, during ourpiloting of the work, the question of whether peopleactually trust the supplier emerged as a key factor.

4. Our approach

Our panel’s attitudesAt the end of the assessment, we asked our panelliststo give us some insight into their attitudes totechnology, personal information and privacy.

The panel want offers by email, not phone.

They strongly agreed with “I always give my emailaddress so that I can get special deals”

They disagreed that “email marketing is a waste oftime because I just delete them “

They didn’t want more offers by phone and, indeed,most of our panellists told us they try to avoid givingphone numbers on web-forms so they aren’tcontacted.

They tend to be ambivalent about advertising as theydidn’t have a strong opinion either way about whetherhaving search paid for by advertising is a good thing.

They aren’t worried about privacy. Most agreed that“worries about giving away too much information areover-stated” and disagreed with “liking the idea aboutgetting offers without giving personal details away”.

How we analysed the feedback

Transparency A good service makes it clear how thedata is collected and used

Access Can you get a copy of the data that isheld about you?

Control A good service allows fine control ofthat data

Trust An assessment of the degree of trustour users felt

Page 8: "Dont target me ask me" ctrl shift report

PAGE 8 © Ctrl-Shift 2012

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

We focused on eight representatives of business modelsaiming to exploit Volunteered Personal Information invarious ways.

Our first group (see methodology, as the assessment didnot run in this order) is based on an advertising basedapproach.

The second uses shopping history to become arecommendation engine (you bought this, so this otherproduct will be relevant).

The third group, the new entrants, aim to createrecommendations based on your own information inputinto their system.

An advertising network that uses your browsing history to serve uprelevant adverts.

The dominant search engine tracksusers and runs on an advertising-based model.

The premier social media site whichhas access to extraordinarily detailedinformation on individuals and uses itto tailor advertising.

The premier online retailer collectsdata on purchases made throughAmazon and responds with tailored offers.

We chose two well known UK-basedbrands (Tesco and Nectar) but somepanellists were members of otherschemes too.

These schemes are, in effect, run byglobal companies, Dunnhumby andAimia, respectively, and so ourfindings can be applied to both theUK and US markets.

4. Our approach CONTINUED >

Advertising

Shopping history

Recommends a set of deals (fromthousands of daily offers) based onpersonal preferences and “all withoutsending you a single email”.

Uses a personality test to drive bettertargeted advertising. For users, theyget a free personality assessment andadvertisers get adverts served thatare more relevant to the specific user groups.

Bases its recommendations on yourreviews of products and services; andshares these with other Hunch users.It takes a crowdsourcing approach(but driven by machines) torecommendations – ‘we’ve got ahunch you’ll like this’ isn’t amarketing strapline of theirs, butcould be.

Personal input

VisualDNATM

Page 9: "Dont target me ask me" ctrl shift report

PAGE 9 © Ctrl-Shift 2012

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

5.1 Inferred interest

Our first grouping of businesses was those with anadvertising-based business model. Of course, each one ofthese explicitly relies on volunteered information –Facebook is all about me; Google search is, by definition,something that I want to know about and Bluekai is simplya real record of websites that I have visited.

First, the overall sense from the panel was that they:

• had little knowledge about what each company held onthem;

• had not looked at it before; and

• were surprised by what they found.

“ I simply didn’t know that this sort of stuff wasbeing collected ” (Bluekai)

There was an underlying suspicion that data put onFacebook was highly vulnerable to misuse and difficult tocontrol. Best not to put it there in the first place theythought.

As consumers, they did not see any benefits to them inFacebook’s use of the information to tailor advertising. Itsattempts were so inaccurate as to be completely ignored.

Google surprised our panel, both with the deductions itmade, sometimes more accurate than others, but alsowith the overall profile they presented. Nevertheless, itwas clear that in many circumstances, our users foundthat information varied stored varied.

“ My work computer had loads of details about me,but my home laptop didn’t ”

And that different people using the same machine causeda core problem of “who am I”?

“ I don’t know if my girlfriend has switched on anyprivacy settings ” (Google)

5. Our findings

Facebook Bluekai Google

Transparency

Access

Control

Trust

Comments on data held and profiles“

“ I worry that information I put on there may be used byother people, but not by Facebook”

“ I don't trust them and limit the information I input. I onlyput information there that I would be comfortable for theentire world to know.”

“ It's not even sure what country I'm in. 'High net worthprofessional' - if only! ”

“ It's inaccurate and it's a bit spooky but I knew my historywas being tracked' sums up my thoughts ”

“ Google seems to think I’m aged 45-54 just because I liketravel, food and politics? That is way out, but theydefinitely got some of my interests right.”

“ It doesn’t bother me as they have got it basically correctand it is kept anonymous, I’m quite surprised at howaccurate it is actually.”

Good PoorOK

Page 10: "Dont target me ask me" ctrl shift report

PAGE 10 © Ctrl-Shift 2012

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

The last observation that came out from our assessment isthe question of timeliness. There was little indication thatbusinesses that infer your interest are able to work outwhat you are currently interested in. For example, one ofthe sites identified that one of our testers was based inthe Netherlands – but that was 10 years ago!

Indeed, there was a sense that “Who am I” was somethingclose to a complete guess.

“ While close to the mark on a few things, it doesn'tdistinguish me from the wife and some of itsassumptions are way, way off ”

“ I can’t work out how Google works out who I am ” (30 year old, whom Google thinks is 45-54)

“ I don’t use Facebook very often so it’s not like me ”

Of course, this doesn’t apply to Facebook pages, sincethese are largely self-generated. However, since thebusiness model of Facebook is to serve up relevant ads,there is clearly a missing link between thedata/information on the page and the inference that thiscan be used to serve relevant ads.

The advertising industry tends to use terms such as click-thru to measure success. However, in many ways thereverse is true. Wastage is legion. For example 9,995people out of 10,000 don't click on Facebook ads. AndGoogle? Well it is forty times better than Facebook (4 in1000) but that still leaves 9960 ads clogging up pages,costing time and money.

We could not help noticing the relationship between ourpanel’s comfort with the data held by a firm and thatfirm’s relative development of its use of personalinformation, and the inclination to respond to advertisingon site.

Facebook, who came out worst on the perception abouthow it manages personal information, also had the worstresponse to advertising. Likewise Google, who performedbest in our survey, also had the most willing respondentsto advertising.

Getting the consumer’s perspective right is clearly asuccess factor in effective marketing, yet we had a senseof these sites serving up content based largely onguesswork about an individual’s profile.

5. Our findings CONTINUED >

Responses to advertising “

“ I see it but ignore it because I’ve “tuned out” of advertseverywhere ”

“ I see it and ignore it because it is not relevant to me ”

“ They think I am only interested in TV and movies, I have alot more interests than that! ”

“ It’s ok because it is anonymous ”

“ I wish it was more accurate so I can get more relevant ads ”

Page 11: "Dont target me ask me" ctrl shift report

PAGE 11 © Ctrl-Shift 2012 Design: Boilerhouse.co.uk

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

5. Our findings CONTINUED >

“What is thatadvert for”

“ A male and a female, interesting.”

“Limited ability to changeyour interests.”

Page 12: "Dont target me ask me" ctrl shift report

PAGE 12 © Ctrl-Shift 2012

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

5.2 Loyalty, retailers and deduced interests

One way that relevant content can be served is to tie itinto specific buying history. For obvious reasons, this is adefining moment. The buyer has clearly shown that theyare more than just interested in certain products.

For this research we chose to look at suppliers who workacross different market sectors because these businessesare generally trying to get their customers to buy moregoods – often in other market sectors. [To put it anotherway, if you join a petrol loyalty scheme, it’s pretty certainthat you are mainly interested in getting rewards forbuying petrol.]

We chose Amazon and then suggested to our panel thatthey think about how they used Tesco’s Clubcard and/orthe Nectar loyalty scheme.

The first point to make is that we aren’t comparing like-with-like simply because the loyalty schemes work largelyin an off-line mode. This means that customers can’t seewhat data is held about them and the only way of judgingwhether or not they “know their customer” is on the basisof the offers and vouchers that are made. Therefore,unlike the other services under assessment, we had to relyon the assessor’s memory.

Loyalty

Loyalty is business nirvana. Yet people who we screenedfor the work, had some interesting perspectives.

“ Is Clubcard a loyalty scheme? ”

“ I’d never use Tesco’s clubcard, I really don’t wantthem knowing all about me ”

“ I can’t see the point of any loyalty card ”

Amazon was more trusted with the information, notbecause of the information quality, which seemed a simplebut accurate list of their purchases, but because Amazonwas a trusted brand. The information they contributed wasexpected to be safe. The advertising that resulted thoughseemed a bit misdirected.

In part this may be because of the persona problem thatwe have already seen,

“ Can I use my husband’s Amazon account as weshare it? ”

“ I use a work colleague’s Amazon account as this ismore convenient ”

but also because of timeliness and relevance.

“ It has no relevance to me today. It is all more than3 years out of date. ”

Returning to our theme of advertising, they key test forthe companies that deduce interests is whether therecommendations offered were taken up. We were slightlysurprised by our findings here. The most frequent responseto “what has been your experience with recommendationon Amazon was “I see them and occasionally click onthem”. For the loyalty schemes, we asked how relevantthe offers and vouchers were and the most frequentresponse was “Partly relevant (I sometimes use thevouchers)”.

It isn’t that loyalty needs to be particularly rewarding, ourpanellists were mostly cashing in their vouchers for £20-£50 over a six month period.

Despite all the data that is being collected by the loyaltyschemes, there was an underlying current….

“ They're clueless.”

“ They aren't very good with the data that theyhave about me.”

In conclusion, therefore, it seems as though reviewingactual purchases probably helps the customer more thanthe retailer. It helps the customer keep a record ofpurchases and to use other people’s recommendations foralternatives but the profile that is built up as a result isclearly not about a single person.

5. Our findings CONTINUED >

Amazon Clubcard/Nectar

Transparency

Access

Control

Trust

Good PoorOK

Page 13: "Dont target me ask me" ctrl shift report

PAGE 13 © Ctrl-Shift 2012 Design: Boilerhouse.co.uk

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

5. Our findings CONTINUED >

“The items bought regularlyare easy to find but nothingabout preferences.”

“Traditional segmentation atplay here - how many kids doyou have”

“Just who do they think thisperson is? Lady Gaga?”

Page 14: "Dont target me ask me" ctrl shift report

PAGE 14 © Ctrl-Shift 2012

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

5.3 Getting customers to volunteer interests

Our next step was to ignore past behaviour as a guide tothe future. Instead, we had three applications that rely onpeople inputting information. The information used iswhat is called non-personally identifiable information –information that only I know but if you have it, you can’twork out it is me.

If you use this effectively, it should reduce worries aboutprivacy and yet make it highly personal. Conceptually, ourpanel understood that in exchange for more personaldetails you might get better deals.

“ Sounds like a great idea ”

But there is a core concern that makes people inherentlysuspicious and questioning whether there is an underlyingmotive.

“ I can’t work out how these companies make money ”

And a certain degree of scepticism about whether ourpanellists were quite ready to use their smartphones.

“ I'd say I'm outside of the key demographic forsomething like this - someone in their early 20swould be much more open to this ” (32 year old)

“ But won’t they know who I am because they’ll beable to work out my phone number and location? ”

VPI-specific business models revolve around the basicpremise that you supply information about yourself, theythen arrange for the supplier to contact you. The contactof choice for suppliers has often been to offer theconsumer a discount, voucher or other financial reward ifthey buy the product being promoted.

Our panel liked this financial reward aspect. It reallyencouraged them to contribute information andinvestigate VPI options after the research.

They also liked the ability to alter the profile to generatemore precisely the offers they wanted. Although thisdidn’t apply to VisualDNA in which you have to re-take thetest as there is no direct link for the user between theimages you select and your personality test. In otherwords, unlike Dealboard where you can select particularaspects to change your profile, you can’t pick outselective parts.

“ Oh I get it now, if I adjust the profile, I can getstuff that is more relevant to me ” (Dealboard)

5. Our findings CONTINUED >

VisualDNA Dealboard Hunch

Transparency

Access

Control

Trust

Comments about personal profiles“

“ Quite like me, slightly wary knowing that it is an advert ”

“ It says I probably live in a household of 5+ people but Ihave no kids - why? Also that I don’t love travel when thatis one of my defining characteristics - I love to travel! ”

“ I quite like the idea that you can select this informationwithout entering personal details ”

“ OK there are no personal details but it relies on myinformation ”

“ This profile was definitely the most like me, I was actuallyquite shocked at how accurate it was after pressing just afew buttons ”

“ Reads like a horoscope, it uses lots of very general phrasesand doesn’t say much ”

VisualDNATM

Good PoorOK

Page 15: "Dont target me ask me" ctrl shift report

PAGE 15 © Ctrl-Shift 2012

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

All of these approaches caused interest from panellists.

“ It’s different, intuitive, clever and made me thinkin a different way.” (about VisualDNA)

Yet one crucial measure – transparency – drew thiscomment:

“ VisualDNA makes a big thing of the personalitytests but then asks you all same stuff thatmarketers ask – age, relationship etc – so I’mworried that it is all a big front for an advertising scam ”

The others did not go without their comments either.Dealboard was found to be very easy to use.

“Dealboard is really easy to use, quick which isimportant and nice as an app ”

It was let down by the narrow range goods offered,however and by the process of sorting the brandpreferences.

“ The problem was that the deals were all aboutlooking good or house stuff ”

Hunch was entertaining, but panellists did not know whatuse it was to them.

“ It seems a bit random ”

What all of these applications did, which none of thetraditional approaches came even close to, was deliver aninstant result. By spending just a few minutes inputtingnon-personally identifiable data got people their profile,some recommendations or some deals.

Overall, how well an application caught the essence of theconsumer and that in turn generated what they wanted,when they wanted it was found to be the benchmark ofsuccess for all parties in this relationship. An accurateprofile increased consumer interest.

And what an interest!

Our panellists were intrigued by these applications andmany unprompted declared they would being tryingdifferent applications after the workshops.

This inspiration was diminished a little by some of theunrecognisable profiling. Or by the presentation of someirrelevant offers. But this was more an amusement ratherthan a deal breaker. The proof would be in the results andbenefits they generated. The sting in the tail is whetherthe personal data is used effectively. At least one person,who like the VisualDNA test, worked out that the reasonfor the test was really “to self-profile” – in other words toput oneself in a marketers sights.

Suppliers in this category do need to worry about getting itright – and fast. Giving people too many options may causeproblems.

“ I did look at the longer version of the Hunch quizto add to the profile but there were too manyquestions ”

“ I stopped telling Hunch about myself after about100 questions ”

“ Each version seems to be as easy to use as possiblebut people's attention spans being shorter thanever it's important to engage quickly. ”

A phone is a personal device but whether the mass marketis quite ready isn’t clear.

“ I tend to use my PC for online shopping, buyingsomething on my phone would seem wrong ”(48 year old)

5. Our findings CONTINUED >

Page 16: "Dont target me ask me" ctrl shift report

PAGE 16 © Ctrl-Shift 2012 Design: Boilerhouse.co.uk

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

5. Our findings CONTINUED >

“Adjusting yourprofile is easyto do.”

“Lots of words to describe you but no feedbackon your interests”

“Uses pictures to find out about youbut no feedback”

VisualDNATM

Page 17: "Dont target me ask me" ctrl shift report

Our three groups of providers went largely asexpected.

1. Those that infer data from browsing habitsdo the worst at using their customers’information.

2. Those that deduce data from "buying" habitsdo better.

3. Those that request non-personallyidentifiable data can do best.

The challenge for the third group is still in ensuring thatthe content served is relevant. And indeed, because thebasis of the profile is based on non-personalised data, theyface a real challenge of serving up the “test” in a relevantway in the first place.

Although initially some were quite sceptical about anyrelationship with a supplier being something they wanted,all were rather attracted to what was on offer in VPI-oriented applications.

The scepticism many started with was multiplied once thefacts about what was known about them already (and howthat information could be used) were made clear. Theywere largely unprepared for the extent of data collectedabout them.

When we asked them to review some of this information,most did not think that it accurately represented who theyare either.

Not off to a good start then?

Quite the contrary. When the personal information basedapplications where introduced and tried, most wereexcited and intrigued by the possible uses.

They were especially attracted by the ones that allowedthem to record accurately in ‘relevant’, what theywanted, terms.

The fact that they contributed the data, or decided whento contribute (ie, turning on or off a service) eased theirunderlying fears about privacy.

It is perhaps a little surprising to some that the sometimesdismissed ‘soulless progress of technology’ in a marketeconomy has been driven, at least in this instance, towarddelivering benefits to both suppliers and buyers. Whowould have thought that suppliers would be willing to givesignificant incentives to consumers if they just stuck theirhand up and said ‘I want one of those now’, savingsuppliers the effort of bombarding them with offers all thetime?

Or that individuals as consumers would welcome theincentives so much that they would be excited abouthelping suppliers find them at the right time?

PAGE 17 © Ctrl-Shift 2012

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

6. Summary results

Total Score Transparency Access Control Trust

Dealboard

Amazon

VisualDNA

Hunch

Clubcard/Nectar

Google

Facebook

Bluekai

10

9

7

7

6

6

5

4

Good PoorOK

Page 18: "Dont target me ask me" ctrl shift report

The challenges we examined at the beginning ofthe research - data; advertising and newtechnology – are now highlighted by the themes emerging.

First there is the question of persona. Most of theexamples we looked at had one flaw – the device orservice wasn’t personal. Searches done at work; sharedlaptops or shared accounts all combine to confoundbusinesses trying to guess what people are truly interested in.

Second, was the question of timeliness. To build up aprofile online (or in a store) takes time (indeed, the term“account history” or “search history” tells its own story).The profile, judging by our panel, can be “quite like me”but that doesn’t necessarily reflect what the user isinterested today.

Finally, was a conundrum, the advance of technology – thesmartphone. This is a “personal” device. Although somepeople have more than one phone (see persona above),it’s unusual event that it might be shared across manypeople. What we heard from our panel were mixedfeelings. They realised that delivery on a phone did makeit personal but weren’t sure that they would do much withthe information.

Three generalisations emerge from this research:

• while incentives matter (it's why you do it in the firstplace), intrinsics also matter. If the process isintrinsically rewarding (i.e. interesting, you arediscovering something about yourself, or it's a bit of achallenge like going through the levels of the game), thismay be crucial to the levels of engagement you achieve.(Ditto, if you make it clunky, frustrating or difficult);

• iteration is a form of addiction. As per game playing, ifyou can see it change in response to what you are doing,it sucks you in. This can be really powerful if theiteration is a way to learn about yourself; and

• trust matters.

On this point, it may be worthwhile pointing out that VPIis a journey along a spectrum. Some staging posts alongthe way include those listed below. The Ctrl-Shifthypothesis is that there is a journey through the spectrumwhere different degrees/levels of information sharingrequire different degrees/levels of trust/personal control.

• ‘Permission to collect'- as in Clubcard and Nectar.

• Volunteered as a component of a broader process, withno visibility or control over how the information issubsequently used.

• Volunteered as a component of a process, with promisesover how the information will be used.

• Volunteered as part of a process with visibility and/orcontrol over how the information will be used (e.g. theTAC of TACT).

• Where the volunteered information is knowingly andintentionally a core element of the value add (as inDealboard)..

The interesting thing with Dealboard and its cohort is therelationship between 'volunteered' and 'inferred'.VisualDNA, Hunch and Dealboard are asking people tovolunteer information in order to make richer, moreaccurate inferences.

This is very different to the Request For Proposals/intentcasting approach: a direct response model.

These approaches will stall unless they become iterative:'we've used the information you volunteered to make thisinference [which is also potentially interesting insight/discovery for individual]. But tell us if we've got somethingwrong, and we will make our inferences even better'.

The future looks bright for the relationship betweentechnology, enterprise and the individual, at least whereVolunteered Personal Information applications restoretrust and balance to the system. It is a relationship thatappeals to both parties. That is why Ctrl-Shift haspredicted that the market for VPI will be worth £20bn inthe UK alone by 2020.

PAGE 18 © Ctrl-Shift 2012

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

7. Conclusions

Page 19: "Dont target me ask me" ctrl shift report

Game-changing

But even beyond this this commercialimperative for businesses to act is a deeperissue. Volunteered personal information willonly be volunteered when customers feelcomfortable that the information is going to beused correctly.

Getting this trust framework right is no small feat – itinvolves up-ending the current relationships. At themoment, all these policies are set by the terms andconditions of the supplier (or by their privacy policy) andthe consumer has no choice in accepting them. However,as other previous Ctrl-Shift research has found, inagreeing these terms, consumers are often forced in tosome form of data-sharing with third parties. This is notthe way to build a better information sharing policy. Afterall, it is only the individual that can decide if they arewilling to share information – privacy is a personal setting.

Critically, these services create a direct, added value linkbetween VPI and traditional marketing analytics(propensity modelling and so on). Before we started thisstudy, we thought VPI would mostly replace analytics.

But with this, it looks like VPI will both enrich analyticsfor businesses and transform analytics into a tool for thecustomer.

That really is game-changing.

PAGE 19 © Ctrl-Shift 2012

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

8. Comment

About Ctrl-ShiftCtrl-Shift is an independent research and consultingbusiness. It has been commissioned for this study but thereport has been produced without the direct involvementof nFluence. It is the strategic advisor to the UKgovernment’s midata project; it is a member of the WorldEconomic Forum’s project and is currently advising anumber of global brands on their personal informationmanagement strategies.

About nFluenceHeadquartered in Seattle, Washington and London,England, nFluence Media is a marketing technologystartup that is focused on making marketing work betterfrom a consumer’s point of view. Their patent-pendingbrand sorter technology allows consumers to createanonymous interest graphs, or “advertars,” for morerelevant targeting of deals, offers, ads andmessages.Mobile carriers, broadcasters, retailers andmedia owners can radically enhance the responses to theirassets with massively reduced intrusion on uninterestedconsumers.

Page 20: "Dont target me ask me" ctrl shift report

www.ctrl-shift.co.uk Don’t target me - ask me (and make it fun)

PAGE 20 © Ctrl-Shift 2012

Ctrl-ShiftSomerset HouseWest Wing, StrandLondon, WC2R 1LA

t| +44 (0) 20 7000 1907e| [email protected]| www.ctrl-shift.co.uk

Design: Boilerhouse.co.uk