7. 2010
(25 exobytes per month)
2011
(32 exobytes per month)
2012
(40 exobytes per month)
2013
(52 exobytes per month)
2014
(65 exobytes per month)
2015
(85 exobytes per month)
From 2012 to 2015 internet traffic
has increased by >100%
In 2012 <0.5% of
data generated was
analysed
Why IOM matters: opportunity
11. “The most profound technologies are
those that disappear. They weave
themselves into the fabric of everyday
life until they are indistinguishable
from it.”
#ProvokeEvent
Mark Weiser, 1991 “The Computer for the 21st Century”
17. “What information consumes is rather
obvious: it consumes the attention of
its recipients.
Hence a wealth of information
creates a poverty of attention.”
Herbert A. Simon, 1971
#ProvokeEvent
Thanks Ian. So today we’re looking at the “Intranet of me”, and exploring what it means to try and create compelling ways of engaging with our customers in an age of constant acceleration.
1. I’m going to reflect on how this next generation of personalisation creates opportunity on an epic scale
2. Then I’m going to unpack the nature of the challenge we face as we try to get value from of all this customer data
3. Throughout I’ll set out some examples of how real-world organisations are responding to these challenges and creating disruption and competitive advantage as a result
[Please tweet using #ProvokeEvent]
So first of all, what do we mean by “The Internet of me?”. As everyday objects and experiences become digitized, we’re seeing new opportunities to interact with customers that are very different from the interactions we’ve seen previously. We can pay for things with online social interactions, our household appliances can learn our behaviour patterns to serve us better, as we travel our experiences can be overlaid with relevant contextual information through wearables and mobile devices. All this demands that as marketers we really understand who and what we’re dealing with…
Now services tailored to the individual used to be the ONLY way we did business until industrialisation and the later rise of supermarkets. In the 1700s, as shop owners we’d have a personal relationship with all our customers. We knew them. We knew what they liked and what would work for them.
Having initially lost that human connection online, businesses started to rely on ways to make our services and products more tailored to individuals…
But why is this important?
The first reason is we’ve seen massive business benefit from aligning our services to customer preferences. [Amazon first iterations of recommendation], and we’re constantly seeing the importance and power of knowing our customers as real individuals. Data shows that people buy more and/or more often when met with personalised retail experiences. In fact more than 75% specifically use websites that cater for their personal preferences based on previous behaviour.
The second reason, is that as humans we crave relationships. And with the rise of social, we expect brands to relate to us in a meaningful way.
Some brands thrive on connecting on a human level rather than about their products.. The equivalent of having a hairdresser who makes small talk with you and makes you feel like you’re a friend.
So I don’t know if you noticed but there was a tube strike last week. Here’s an example of this taking a topics that’s removed from their product offering and putting their own unique spin on it.
Crucially, they also talk directly to individuals, making the experience even more individual.
The final reason IOM matters is the scale of the opportunity it represents. The conversation is getting more and more intense.
If we understand what a gigabyte is, then an exabyte is one billion gigabytes
In Dec 2012, IDC estimated that <0.5% of data captured is ever analysed (area of circle)
In fact the growth between 2014 2015 is as big as all internet traffic for 2010! (Cisco)
Mostly unstructured data, such as messaging, social posts, images and video.
It’s only accelerating. By about 2020, the planned Square Kilometer Array (SKA) telescope alone will generate 1 Exabyte of data every day!
If we look at where that acceleration is coming from, it’s not from people sat in front of a laptop. It’s people out walking their dogs. It’s the dogs being walked in some cases, with internet enabled collars (a kind of MapMyHound). It’s our cars and houses and TVs and watches having micro-interactions that build up an amazing digital picture of our connected lives.
QUESTION: Who here owns a smartwatch or smartband? Who here owns a tablet? Who owns a smartphone? Who 5 years ago owned a tablet? Who ten years ago owned a smartphone? Adoption waves it’s coming and we need to be ready
Wearables are a key part of the internet of me, because they provide an amazing amount of context and information that can be used to create an individual experience. We’ve still got a long way to go… [Notifications across devices story]
Although we’re early on in the journey of wearables and personalisation, we can get an idea of the power of this from what Disney have demonstrated with their Magic Band. Last year they introduced personalisation driven by Magic Band data, so your child’s favourite character could come up and greet them by name and ask them about the rides it knows they’ve been. You can get tailored offers and services based on where you are and your known preferences and so on, as well as the incredible data it provides back to Disney on waiting times, behaviour patterns, sales and so on.
What Disney have appreciated is that the technology is not the important thing. [Quote] The important thing is the experience that is created when the technology just invisibly does what’s necessary to enable it.
Two more quick examples of how travel brands have taken technology and made that an invisible part of a great individual experience. KLM Surprise have tracked people using a variety of social channels, then given personalised gifts at check-in based on what it knows about them. They’ve also used the same social tracking to suggest seat choices based on who KLM think you’ll get along with. The second example is Virgin Atlantic who generated a lot of publicity last year for using Google glass to identify customers and improve their concierge services, being able to know who the customer is and what their needs are likely to be. No evidence that this was much more than a gimmick, but companies will continue to explore and test opportunities like this.
A more familiar example of individual differences being weaved into the fabric of life is Google search. Two people will get very different search results depending on what else Google knows about them to provide additional context or opportunities.
It’s easy then to project this and create a utopian vision where my wearable devices offer me discounts on treats when I’ve done more than my daily recommended exercise threshold. My car communicates seamlessly with traffic systems to optimise flow and improve safety. My insurance company reduces my premiums based on health data and my connected toothbrush. Everything connects to give me a tailor-made digital experience.
But before we get too carried away, there are three significant mountains we need to scale to take advantage of this.
The first is the volume and speed of data being thrown at you. We’re used to dealing with a handful of data points (is this customer new or returning, does this customer fall into one of our pre-defined groups, …) but the data being generated from wearables and IoT (my driving habits, the patterns of energy and appliance use at home, and my devices talking to my other devices) means we’re standing in front of a firehose of data.
Nobel laureate economist and a founding father of AI Herbert A Simon recognised the challenge this creates. [QUOTE] Our businesses will struggle to deal with the wealth of information unless we’re able to find the treasure within it. This leads us on to our second challenge…
The second is that we often think about mining data for some kind of gold nuggets of meaning. The reality is that in this new world we’re just creating an ocean of data, and it’s impossible to know where the value lies or what it looks like. So the challenge is to create ways of looking at the data without being 100% sure what we’re going to use the data for. This raises the challenge of creating effective systems for machine learning to drive this.
One powerful outcome of machine learning is in finding non-obvious relationships in unstructured data. As an example a really good indicator of people being a good credit risk is people who at some point have bought anti-scuff pads for their furniture. Unlikely to have been spotted by underwriters, but obvious to a machine who doesn’t approach things in the same way as us. However, we’re in the infancy of machine learning on the scale that’s necessary for us to look at >0.5% of the data we’re collecting.
Finally, we have the complexity introduced by millennials. Now this group already expect the world to revolve around them, and assume that because they can have a great experience on one platform in one channel it should just work across everything else, right?
However as we add more complexity and relationships between data that we need to take account of, we rapidly increase the number of combinations that we need to be analysing to truly create a market of one…
As a simple analogy, the complexity of the cube on the right (6x6x6)is not twice as much as the one on the left(3x3x3). It’s not even twice cubed (8) times as complex. In fact it’s more complicated to the tune of a septennonagintillion, or 1 with 97 zeros after it, or to put it another way, if the complexity of the normal rubix cube is an atom, the complexity of the version on the right is more than the number of atoms in the entire universe. That’s how difficult we make things when we start to add in many different parameters and try and get meaning from it.
In light of this we come back to the rise of the machines…
..and it raises an existential question. What’s the future for marketers? What’s the implication if your digital channels are spotting trends without you and updating themselves in realtime to better target individuals?
Machine learning means systems will get better at testing and predicting ways to optimise online solutions, and ultimately to keep up with the level and complexity our customers will expect, we will have to give over more and more of the manual reviewing and decision-making to machines.
So the crucial question is how will you continue to demonstrate value?
For me it’s all about looking for the things that people do really, really well and machines don’t. For example, machines are terrible at knowing what questions we should be asking?
Creating art & stories – these are compelling and intrinsically human
Flexibility
Unstructured problem solving
Identifying something relevant in a flood of undefined phenomena
Expressing empathy
Making us laugh
Ability to break the rules
To close I wanted to talk about three examples where the internet of me is starting to become more visible.
The first is that when transport in London is actually running, beacons will allow advertisers to push targeted offers and content to commuters. With average journey times of 15-20 minutes on a bus, it’s seen as a significant opportunity to push personalised offers based on profiles, location and so on.
In the past few years machine learning has been used to complement crowd sourced solutions for identifying things which share common characteristics (e.g. automation of the GalaxyZoo project)…
..but now the principles have been proven, they are being applied to identify subtle emotional differences detected on webcams to help marketers track reactions and attention levels, then be able to change content in order to improve conversions. (Realeyes)
Slightly creepy? Yes. But we’re willing participants when we opt in to schemes like this in return for incentives and discounts.
As a final example of this, Southwest airlines are using machine learning with speech analytics to assess live-recorded interactions between customers and personnel to provide big-data insights on how they can serve their customers better and get better outcomes with subtle changes in tone. Speech analytics presents a really powerful avenue as so much data contains voice elements.
One common theme across provoke is how we can all ‘hack the enterprise’ to prepare us for the perfect storm of disruption that we’re entering into. As such we need to find low cost, low barriers to entry ways of applying principles to gain benefit.
As such, my three takeaways would be:
Start to think about what a brilliant experience would be like – ignoring any technical constraints. Chances are someone else will break those technical constraints for you by the time you’re ready to use it
Don’t ignore the IOT. It’s where the majority of device and data growth will be happening over the coming years. Are there simple ways to prototype new ways of gaining insights that will help you give better service to customers
Don’t wait for you role to be redefined for you. Take control and look for ways that machine learning can free you up to be a more human marketer!
Start to think about what a brilliant experience would be like – ignoring any technical constraints. Chances are someone else will break those technical constraints for you by the time you’re ready to use it
Don’t ignore the IOT. It’s where the majority of device and data growth will be happening over the coming years. Are there simple ways to prototype new ways of gaining insights that will help you give better service to customers
Don’t wait for you role to be redefined for you. Take control and look for ways that machine learning can free you up to be a more