The cookie conundrum

Over the past 12 months there has been considerable amount of press around cookies and tracking of users with the likes of Microsoft and Mozilla blocking unrelated 3rd parties from tracking users as they browse and tools blocking altogether. 

While there needs to be a move away from the wild west that went before us (which will ultimately benefit the consumer and advertisers alike) some of the current thinking does need to be refined.

We also have to consider that when used correctly cookies can aid with the ‘operational’ side of browsing, making repeat visits to commonly used websites such as banking a far more seamless – so blocking "all" really is not an option. 

And for advertisers the stakes are high with the alternatives to cookies being potentially more intrusive, one such alternative is device fingerprinting. 

Device fingerprinting

To uniquely identify a person 12 points on a finger print are needed – to uniquely identify a device such as your mobile far fewer are required.  Each device has a unique finger print, even same models and manufacturers – web browser, plugins, language settings, IP address of the device when all stitched together can be used by an advertiser to identify you and your device.

And it does not stop there, merging seemingly innocent data sources together can in some cases reveal more then you could imagine. As an example MIT labs were able to reliably identify a user from 1.5 million anonymous cell phone registrations and some twitter data.  Remember, companies such as Apple collect anonymous data for their and other third parties companies use with our consent (each click of those terms and conditions enters our data into that pot) which if used in the wrong way could have dire consequences.

So how could the future look? 

Where ‘big data’ is concerned there are suggestions organisations are the stewards of user data, but are they truly effective at this task?  As an organisation keeping track of both what you are holding, where it was obtained, the usage rights around that data and most importantly when it should be deleted is beyond all but the most advanced companies.

Putting the user back in control

As a user I’m ok with sharing my data for certain uses which are beneficial to me, other times not. The current method of opting in is too simple and does not consider the context in which my data is going to be used now or in the future.

Putting the user back in control of the data may sound daunting, however the alternative of simply blocking everything or having to opt in may quite simply be too basic.

One concept that in the future may hold some promise is the idea of a personal data store. 

Simply put all your data is stored securely in a virtual locker – applications then request access to the data they need, the user at the point of request can either allow or deny access as they deem appropriate and at a later date be able to change.  In a world where transactions are occurring machine to machine, sometimes without our knowledge gaining visibility and clarity can only be a good thing.

Whatever the answer any organisation that manages to earn the trust of it’s users, be clear on what data is being collected, how it’s going to be used, gives easy opt out options and compensation (directly or through some form of value exchange) and is able to leverage data for both it’s own and it’s users advantage will win in the end.


 Law 1: Reduce
The simplest way to achieve simplicity is through thoughtful reduction.

Law 2: Organize
Organization makes a system of many appear fewer.

Law 3: Time
Savings in time feel like simplicity.

Law 4: Learn
Knowledge makes everything simpler.

Law 5: Differences
Simplicity and complexity need each other.

Law 6: Context
What lies in the periphery of simplicity is definitely not peripheral.

Law 7: Emotion
More emotions are better than less.

Law 8: Trust
In simplicity we trust.

Law 9: Failure

Some things can never be made simple.

Law 10: The One
Simplicity is about subtracting the obvious, and adding the meaningful.


“Engagement” – are you paying attention?

Microsoft’s new home gaming console – the X-Box One again pushes the boundaries of gaming entertainment combined with the all-important Kinect motion sensor.

Since launch, the Kinect has been hacked to self-drive a mini car to powering a real time light sabre; use cases beyond anything MSFT had imagined when they conceived the device

Now with new enhanced compute power and an upgraded Kinect sensor, what does the future hold beyond realistic engaging content and game play?

The new Kinetic technical specification are impressive – with increased field view, infrared and HD cameras allowing it to build up better 3D images of the room than before, leading to exciting possibilities in both gaming and marketing.

The ability to measure heart rate through the fluctuations in skin tone combined with your pupil size could be used to measure your level of engagement with a game or advert, changing either the difficulty or brand message accordingly.

Multiple people in the room, who is actually paying attention to that advert? 
Body temperatures rising? – How about a quick advert for an ice cold Coke …. along with a family pizza meal deal as there are 4 of you in the room.

Or even worse, working out when you are tired (and more likely to make impulse purchases) to hit you with some brand messages….

Understanding the context of the signals is critical; as certain signals could have dual meaning and if interpreted incorrectly have dire consequences - tears of joy / tears of despair.

Conscious of this along with just the general creep factor around being watched and the actions that this spawns it still remains an exciting opportunity to really get ‘engaged’


Big Data - keeping it real

Just this week we spent a great couple of days in Chicago at the Pass Business Analytics conference, presenting our thoughts and technology around digital attribution. 
In the space of 'Big Data' scale does have its advantages – especially in modelling with the ability to keep those "outliers" in the data set, however there are key principles that still apply when working with 'Big Data' in marketing.
Using the 80/20 rule
20% of the data giving 80% of the value.   
This philosophy I believe should never change, just because the technology allows us sometimes to reverse this rule
Engaging early with the business
Ensuring KPI’s and metrics are agreed and quantified up front – simply thinking you can define these after the event because you are capturing all activity will only lead to complexity, cost and challenges further on.
Causation VS Correlation
They are different.  You must have the context and understanding.
Defining Insights that are actionable
This goes back to my point of early engagement with the business for this to be possible.
Data is after all just data – it is the insights that are derived from this data which is key.
Getting the data in, getting it analysed to make meaningful recommendations in a timely fashion is critical, especially when the window of opportunity is small – we are looking for many incremental improvements, not big bang.


“Thunderstorms” and “Curly Fries” - you're smart

What we share and reveal about ourselves, either consciously through social media or through our browsing habits all contributes to our digital fingerprint – a fingerprint that is then used to categorise us into ‘buckets’ for targeting of marketing activity.
How this fingerprint is constructed is always under scrutiny – with a balance of user privacy and the needs of the publisher to generate revenue from their content in constant tension.
Recently the announcements from Mozilla around 3rd party cookies has refocused people on this topic and how the necessary profiling of people to deliver relevant messages needs to move on from the current methods.
Better use of ‘public’ data to infer gender, brand preference and age is one area of opportunity – but really what can you infer from this freely available information – quite a lot it would seem.
Recent research from the University of Cambridge and Microsoft using Facebook ‘likes’ has been able to correlate ‘likes’ to attributes such as political and sexual orientation, as well as intellect and gender.
Using data from 58,000 volunteers who provided their Facebook Likes, detailed demographic profiles, and the results of several psychometric tests to build models analysts were able predict with high probability which ‘buckets’.
For example, users who “liked” Thunderstorms and Curly fries were predicted to be of higher intelligence than those who showed affinity towards Harley Davidson and Sephora.
On its own it’s application is limited, however if used in combination with others methods of tracking, such as context / semantics and device finger-printing may actually offer a more robust solution (over cookies) to the whole challenge of delivering relevant marketing to users.

research : here