This document discusses the importance of personalization and understanding audiences on a deeper level. It argues that traditional segmentation methods are not enough and personalization should go beyond demographics. Effective personalization requires understanding motivations, needs, values, attitudes and behaviors of audiences through qualitative research. It also stresses that companies should view their offerings from the perspective of the audience and understand how the offerings fit into audiences' lives. Taxonomies need to be as deep as the understanding of audiences and content should supplement discussions important to audiences.
2. A little about us
Reading Room is a global digital communications consultancy,
founded in 1997, independent and owner-operated with over 220
staff passionate about digital in 7 offices serving 4 continents,
Follow us!
@ReadingRoom
www.readingroom.com
3. Dr. Deborah Ko
Behavioural Psychologist
Deborah Ko received her PhD in
psychology with a specialization in the
cultural psychology and the psychology of
Internet behavior.
She currently is a psychology consultant
looking at the psychological influences in
digital, such as its impact in design, user
experience, and content creation. Her
goals are to encourage behavioral change
through digital strategy and
implementation.
Deborah.Ko@readingroom.com.sg //@kesatuuli
4. We have short attention spans.
We don’t like wasting time sifting
through useless information.
We hate making lots of decisions.
7. 1. How old is she?
2. What’s her occupation?
3. Budget?
4. What features?
5. What are her interests?
6. Frustrations about her skin?
7. What makes her think about
skincare?
8. Feelings about youth/aging?
9. What does her appearance say
about her?
10. UK mothers find
marketing messages
overly idealistic and
unachievable
46%
UK mums find
messages sexist.
28%
UK mums actually
identify with
advertisements’
portrayal of mothers.
19%
UK mothers
believe that
marketing
messages view
motherhood “as
an activity” not a
bond between
mother and child
58%
11. South East Asian
mums increase
their Internet usage
above other media
sources once they
become mothers
80% South East Asian
mums prefer
internet
resources for
parenting advice
over traditional
means
74%
Many Chinese women do not relate
to the supermom stereotype,
instead demanding a perspective of
motherhood “that solves real-life
problems”
12. How we typically
segment
Stay at home
With young
children
Work part-time
With teenagers Career driven
With young
adults
With infants
17. Make your
research smarter
Business as usual…
Focus groups to
get opinions from
your audience
Asking people
what they want
Asking them only
about how they
feel about your
product/service
Asking if they
would participate
or like
hypothetical
situations/campai
gns
Asking them
advice on what
would make
things better
BAD FORM.
19. What causes them
to think about
things related to
your industry?
What intrigues
them? What’s
worth sharing
and why?
What scares them?
What keeps them
up at night? How does your
product fit into their
lifestyle?
What
problems
do they
have?
What gives
them a
sense of
pride?
Make your research smarter
Ask lifestyle questions:
21. Psychographic personas
Personas that outline:
MotivationsNeeds Values
Attitudes/Emotions
Barriers/Fears
Behaviours that occur because of the above
22. Although born in Australia, he
moved his wife Sophia, an
engineer, and his son Andrew (age
3) to Singapore to be closer to his
parents who live in Jurong East.
He is an established brand designer,
working at Ogilvy and enjoys working
on elevating local brands.
Although he lives frugally, he likes to
spend on quality products and
experiences. He recently spent a
weekend researching information on a
new guitar since his band plays monthly
at Full Pint Brewery along the East
Coast park.
When he travels he likes to immerse
himself in the culture. Although his
backpacking days are over, he still
enjoys taking the road less travelled,
trying new foods, and picking up
meaningful souvenirs from his travels.
Dissecting the average persona
Travel Persona
David Lim, 34 years old
24. Joking or banter on
forums and private
messages
Meeting new friends that
share your similar values or
experiences
Deep friendships and
offline connections.
Emotional healing from
traumatic events and
lifelines.
Less emotional
connection
High emotional
connection
The Blood Brothers (and Sisters)
Veterans that are looking for community and the ability to connect on an
emotional level with someone that understands them.
It can be to pass the time and have fun or be a lifesaver in darker times.
The forums, private messages, and chat features are most important to them.
26. The Networkers
Veterans that are connecting to build a social network of people they know. They use
the site to find past acquaintances and friends.
The search functions are the most important to them.
Career Personal
From my
unit
From my
friends
From my
training camp
From my
military branch
29. Emotional
appeal
Self
motivation
Functional &
Informational
knowledge
Content and the consumer mind
How useful is the solution? How
difficult is it for me to carry out
the solution? Do I have the
necessary mental and physical
capacity to carry out the
solution?
How does the problem affect me
emotionally? How would I feel if
the problem was solved? How
assured and supported do I feel
in my decision?
What is the person I want to be?
Who do I want to be perceived
as? What are my values and
attitudes?
32. Offline
Online
Problem | Solution | Research | Decision | Purchase | Validation
Emotions: Hope, anticipation, skepticism, curiosity
Friends/Famil
y
Celebrities Experts
TV
Magazine/N
ewspaper
Blogs
Social
Media
Forums
Web
article
2. Do I know how to fix it? Am I capable?
Info Emo Self
33. Offline
Online
Problem | Solution | Research | Decision | Purchase | Validation
Emotions: Uncertainty, curiosity, fear of regret
3. I think I know what I need. What is the best to
meet my needs?
Friends/Famil
y
Celebrities Experts
Blogs
Social
Media
Forums
Web
report
Sales
person
SMS/
call
Magazine/N
ewspaper
In-store
comparison
Online
product
comparison/
reviews
Online
videos
Blogs
Web-
site
Info Emo Self
34. Offline
Online
Problem | Solution | Research | Decision | Purchase | Validation
Emotions: Hope, anticipation, reassurance, fear, worry
4. I have enough information and confidence to
make a decision
SELF
Effective
Within
budget
Social needs
met
Utility
needs met
Product is
self
expressive
Quality
needs met
Info Emo Self
35. Offline
Online
Problem | Solution | Research | Decision | Purchase | Validation
Emotions: Excitement, frustration, impatience
5. How easy/convenient is it for me to make the
purchase?
Customer
service
Fast
delivery
options
Ease of
checkout
Secure
Sales
person
Online
product
comparison/
reviews
Availability
Info Emo Self
36. Offline
Online
Problem | Solution | Research | Decision | Purchase | Validation
Emotions: Satisfaction, delight, regret, frustration, anger, disappointment
6. Did my choice help solve my problem?
Friends/Famil
y
Celebrities ExpertsBlogs
Social
Media
Forums
Web
press/vide
o
SMS/
call
Advertisers
TV
Magazine/N
ewspaper
Instagram
Online
product
comparison/
reviews
TV
Magazine/New
spaper
eDM
Info Emo Self
37. Bottom line Know what data you want to
gather and what it should
measure
Understand your audience’s
lifestyle and motivations
Segment on a psychological level
Adjust with the needs they have
in their decision making process
40. Revisit your offerings
within MY lifestyle
How does your
product help me
reach my goals in
life? Does it
reflect my own
interests and
values?
When am I going
to be interested
in using your
product? When
would it solve my
problems?
Why is your
product relevant to
my life? How does
your product help
me maintain or
improve my
lifestyle?
Will your product
help me continue
to be happy? Will
your product
help me to avoid
becoming
unhappy?
1 2 3 4
41. Stop talking
about yourself.
Content should supplement a
larger discussion for
engagement
Provide content that provides
value (emotional, functional,
self)
43. 75% of its subscribers decide what to
watch based on Netflix
recommendations
It has 76,897 distinct movie categories
"People consume more hours of video
and stick with the service longer
when we use these tags.” – Todd Yellin,
VP of product innovation, creator of the
Netflix taxonomy
45. "We were doing a
really good job
with mathematical
crunching, but we
needed to know
our content better.
That requires real
humans.”
Todd Yellin
Revisit your taxonomy
Create psychographic and deep-
level tags (humans)
Create tags based on categories
your audience uses to evaluate
content (humans)
Identify the behaviours important
in how users consume your
content (humans)
Personalise based on the
intersection of behavioural
patterns and deep-level tags
(automated)
47. More successful articles had more
humanizing sentences in them
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
1.9 2 2.1 2.2 2.3 2.4 2.5
Percentageofsentenceswith
humanizingwords
Weighted rank of stories
8/7/2015 Architects of digital change 47
There was a positive relationship between success of the article and how many times a
comparison was made between the subject and the reader (i.e., “whether their lives are really so
different from our own”, “we all have the same colour spine”)
Data obtained for June 20 – July 20, 2014 from Google Analytics for top 5 and bottom 5 stories
48. Taxonomy data can be predictive
Good taxonomy is
insightful
Netflix chose to
produce House of
Cards based on
knowing people’s
habits viewing David
Fincher and Kevin
Spacey films, and how
well similar genres
had done (House of
Cards, UK original
series)
Taxonomy helps target
marketing
Kevin Spacey fans saw ads with Kevin
Spacey
“Thelma and Louise” fans saw trailers
showcasing female characters
Film connoisseurs saw trailers that
focused on David Fincher’s directing
49. Bottom line Content direction is fueled by
knowing your audiences on a
MUCH deeper level
Your message and your offerings
are expected to be part of a
broader dialogue in your
audiences’ lives
Your taxonomy needs to be just
as deep as how you know your
audience
50. Understand
my needs
through time
Ask the right
question
What drives me to do something
(motivation)? What stops me from reaching
my goals (hurdles)? What makes me refocus
on my goals (triggers)?
How do I want to be seen (self)? How does it
make me feel (emotional)? Am I
knowledgeable enough about the
problem/solution (functional)?
How does the my question change
through the decision process?
Where do I find the information I need
during the decision process?
So what? – KNOW ME
51. Reevaluate your
current
content/offerings
The solution has
to be based on
humans, not on
algorithms
Categorise your content/products/services
with labels that make sense to your user
(they don’t even need to be visible to your
user)
Identify behaviours that will be measured
and what they mean
Make these categories help you learn about
your audience on a deeper level
Why should I care about what you have
to offer me?
Do you have a place in the dialogue I’m
already having?
So what? – KNOW YOUR PRODUCT
54. Content affects how we problem solve
In work we did with a housing development company in the UK, the way content
was framed momentarily changed how people believed they approached problem
solving.
Editor's Notes
I’m from Reading Room, we’re a digital communications consultancy, originally from London and one of the top UX companies in Singapore. I’m part of their specialist consultants group,
And I look at human interaction with technology and digital communication with a specialisation in APAC. I like to tell people that I’m the bad type of doctor, I can’t prescribe you medication, I can’t save your life, and I can’t explain why your boss is crazy.
What do we know about human beings? Digital or not, we hate wasting time, we have shorter attention spans than ever, and we hate making decisions.
In fact, in a study with doctors in a scenario of either taking medication or hip replacement, they found that if doctors were presented with easy decisions and many decisions, they’d take the easy decision even if it might have been worse for the patient. Another study showed that in presenting consumers with 24 jams versus 6 jams, only 12% will buy a jam out of the 24 versus 48% buying a jam when presented with 6 jams.
We know that digital gives us unprecedented access to information, and presents options and arrays we may never have been provided with before.
So people will talk about personalisation being more helpful in speaking to your users, and making your brand more intimate with user needs, but WHY is it so effective actually?
It’s because personalisation presents us with relevant information that’s effortless, time efficient, and simple. It helps us cut through all the crap that’s now bombarding us on our phones, our tablets, computers, and now our watches. We’re more likely to stay, we are more likely to think positively about our experience with a brand, purely because we didn’t have to think too hard. Purely because we waste less time and accomplish what we want.
So let’s start off with an example and see how personalisation would apply here. What things would we need to know about her in order to personalise to her? What would we need to know in order to actually speak to her? So write down or think about maybe 10 crucial questions you’d like to ask her in the next few seconds.
What questions did you want to ask her? How many had questions like 1, 2, or 3? How about 4 or 5?
Typically, when we think of personalisation, we think of these top 4 things. If we’re really good, we’ll also throw in the 5th one. This is how we target people, and I’ll be honest, it’s easier this way. But if we want to get an edge in terms of really understanding our customers, we need to go a bit deeper. We need to explore 6 – 9.
Part of personalisation means knowing your audience and knowing them well. We usually do this by targeting specific types of audience members – we segment them into meaningful parts. What I’m hear to tell you is that business as usual segmentation isn’t working. If we want true personalisation the way that our users expect it, we have to go much farther than we are now. So I showed you an example of buying wrinkle cream. Fair enough, but let’s take a real world example of why it’s important to go deeper in our understanding of our audience.
A good example is the segment of “mothers.” How do we market to “mothers?” Well, it turns out that “mothers” as we’ve segmented them, isn’t working. From a whole slew of studies, both in the UK, China, and South East Asia, it’s clear that “mothers” do not identify with and are sometimes offended by the way that mothers are portrayed. Marketers are marketing something that real mothers do not identify with. Why is this?
http://www.marketingmagazine.co.uk/article/1344066/why-marketing-mums-broken
http://www.theguardian.com/media-network/marketing-agencies-association-partner-zone/2014/sep/04/truth-marketing-mums-uk-advertising
http://www.stylus.com/gctlgm
http://www.digitalmarket.asia/asian-moms-primary-home-decision-makers-are-going-mobile-first/
http://www.ogilvy.com/News/Press-Releases/March-2010-Mums-the-Word-a-New-Study-from-Ogilvy-China.aspx
So this is business as usual. A marketer will say “we need to market to moms.” We can segment them by their children or their home life – this is all a mother cares about anyway right? This is what defines mothers as mothers.
The real question then becomes – do mothers actually segment themselves this way? Do they only see themselves as caretakers of their children? Do they only see themselves as dominating the home ground or juggling a career?
And of course, the answer is no. This is merely part of how mothers see themselves in their new role.
This is part of the reason that mother blogs are so popular, because the blogs are segmented by much more nuanced topics. In fact, a study of 900 moms found that 81% read more than 5 blogs a week, 84% of those reading blogs based on authentic content that interests them. So yes, a mother may consider her career choices and yes, she will have specific needs related to how she approaches parenting as her child or children develop. But she still has other interests that also are important that are influential to or change because of motherhood. We have interests here at the top, you go further down and maybe it’s emotions and attitudes, and at the bottom perhaps how her identity has changed.
Why do mothers trust other mothers? Why do mothers trust blogging mothers? Because they are real people with complicated lives and interests that are similar to their own. Marketers don’t provide them with someone they find to be real or relatable. So if marketers want to enter in the conversation, they need to throw away their 2 dimensional models because that’s not how people are.
http://www.newmediaandmarketing.com/81-of-moms-read-5-blogs-per-week/
Segmentation has gone from demographics and simple levels, to deep and nuanced levels. It should be multidimensional and show different and important facets of how your audiences see themselves, it should be psychological and reveal their fears, desires, and motivations. It should be holistic and encompass an overview of their lifestyle.
So today I want to talk about ways in which you can start to know your audience better and then how you can use that knowledge to find out how to better understand, evaluate, and create your content.
It starts with research. Most research – how we normally do it – is we get a whole bunch of people into a room and we ask them what they want. We ask them opinions on a product you have or want to create. If we created said product or went with said campaign – would you like it? What features in this product would make it better?
I’m not saying these questions might not be helpful at some point. But it’s not what a user’s job is supposed to be. We hire marketers to understand the best ways to engage an audience they know a lot about. We don’t hire marketers to ask people they don’t know a lot about to tell them how they need to be engaged with. It’s like asking your wife what she’d like for her birthday. You could do that and get her what she says she wants. But what she wants is to not have to be asked. SHE WANTS YOU TO KNOW.
So if you go with business as usual, you basically know how they feel, sitting in a room, with this random product in their hands. The problem with this line of questioning is that you don’t really know what they’re like. You know how they perceive the box right now. But how would they actually use it in their real lives?
Like it or not, human beings are horrible at foretelling how they’ll actually feel about events or things. Studies have shown that participants asked to describe how they’d feel in different scenarios, all participants were less accurate the more in the future the scenarios were. Not only were they less accurate about how’d they feel, they also overestimated the extent to these feelings. For instance, “how happy will you be if we create this product that you’re holding?” do you see where the danger is on relying on people’s assumptions about how they think they’ll feel or what they think they’ll need?
http://www.psych.nyu.edu/trope/Liberman%20et%20al.,%202002%20-%20JESP.pdf
So how do we change what we know? We have to change the type of questions we ask our users. Our objective shouldn’t be how they feel about whatever it is you’re trying to sell them. Your objective is to understand what their lifestyle is like. What goes on in their head when they face hard decisions? And this doesn’t mean you can ask them point blank. It means that you have to start getting creative about how you get information out. For instance, maybe you need to know how they’ve reacted in a variety of situations. Maybe you need to see what other things they’ve already purchased. What events have they been really stressed out in?
Research should tell you what is going on in here. What makes them tick? What surprises them and delights them. What makes them angry or scared? What are their interests? What do they spend their most time on? What do they spend their most time dreaming about? Your job as a marketer is to understand that world they’ve created, and find out how your product or service fits into that world. If you can’t figure it out – you’re not doing your job. It’s not their job, it’s yours.
So once you’ve delved deeper into your audience, you can truly start to segment them on richer details. I call these psychographic personas. They still may have demographic elements, like they might have a certain profession, or be of a certain socioeconomic status. If gender is important, then they might be a certain gender. The thing is, the details that are NOT important to the persona shouldn’t be included. So I shouldn’t tell you it’s a girl if being a girl is not important. You should be designing for a boy or a girl. But I need to be able to tell you that this persona is driven by a need to be seen, whereas this persona is driven by a need for assurance and safety. These are motivations that change how they process information, what things they find important, and what they will best react to. These are the things that you need to understand when you are creating content, communicating, designing, and building.
Born elsewhere – is he actually an expat?
How important is family?
Is it important that he focuses on local brands? Does that say something about his character?
Is it important that he’s working for a large agency like Ogilvy?
Is his musical interest an indicator for a greater artistic passion?
He was a backpacker – does he not care for luxury now?
He spent a whole weekend just on researching a purchase, does he do the same for travel or only music?
He seems adventurous since he backpacked and tries new foods, does he want the same in a hotel?
There’s a lot of clues here and a lot of conclusions I can make, but are they good conclusions?
This is where we need to add in psychological elements into the persona. Instead of telling us what he is doing and focusing on the behaviours, we need to understand WHY he is doing them. What is driving his behaviour? We start with motivations and drives, what moves people towards their goals and aspirations? What are his attitudes and emotions towards his goal? Then we need to look at the hurdles or the barriers – are there things that cause him to be put off by something? What are things that stop him from achieving his goals? And then we should look at triggers, so in all the things that he thinks about on a daily basis, what catches his attention enough to change his thinking and turn his attention towards this particular goal?
Let’s take an example of some personas we did for a military social networking site. We were tasked with finding out if they needed a mobile app. Business as usual would say that we should get the veterans all in a room and ask them if they had a mobile app what things they would like in it, what features they would use. But this shouldn’t be business as usual – we should know what they need. We took a step back to look at why they used it. Military veterans are not all the same. They have very different motivations for going to the website and it can’t easily be boxed into the military branch they were in or how long they had been veterans. We had to segment them through their different motivations.
I’ll go over 2 of the 3 personas we identified. We found this persona, blood brothers and sisters to be individuals that went to the site to be part of a community. They had strong emotional needs and were motivated to find social support with like-minded people who understood the experiences they went through. It could be just from jokes and chats with people that “got military life” all the way to finding others who were suffering from PTSD.
We then identified the motivations they had – towards social support from like-minded people, their attitudes and emotions that might trigger them to think about reconnecting, such as the belief that military veterans are your family, or feeling isolated or lonely, and then we looked at the behaviours that came about because of these motivations and attitudes such as going onto forums and chatting with new found friends. Then we looked at how this fits into their mobile needs – what mobile features would help support this lifestyle? They’d need optimised forum functions, ability to create group chats, notifications on forum activity.
Contrast this with the second persona we created. This group of individuals go on the website specifically to connect with people they know, either to boost their career and keep professional communication in a military setting, or to reconnect with friends that they served with during their time in the military.
When we look at their motivations, it’s very different – they want to reconnect with familiar faces. The attitudes around this is that the military creates ties that they can tap into, they are emotionally triggered by feelings of curiosity “What’s this guy up to now?” and nostalgia “Remember the time when…” and thus their behaviours are very different. They are searching for specific people and then contacting them directly. They don’t care to go on forums because they don’t need to meet new people, just reconnect with old friends. So their mobile needs are different. They need optimised search functionalities, ability for quick private chats and notifications on private messages.
Notice that they are all veterans. But they fall into distinct psychological categories that in turn, can explain very different behaviours, and helps us to understand what product we need to create for them. We can create a mobile app, but to personalise it to these audiences, we need to know about what it means to connect on that site first.
Another important factor is that individuals needs and motivations change over time. In order to personalise affectively, we need to be able to develop with them.
How do consumers decide to engage with your products or services?
These are the types of questions people think about when they seek information, when they evaluate your product or service, if they want to decide your product or service is important to them.
This is the importance of user research. We need to understand not only how they would answer these questions, but more importantly, how your brand helps them achieve these things.
We need to understand their lifestyle and how our products fit into them. What part of their busy life are we satisfying? How does our product or service fit into their existing ecosystem of things and events?
So lets go back to our first example about this girl who we want to be interested in this wrinkle cream.
notice when she first sees a need for an anti-aging product, she’s probably thinking about how she appears to others, what she’d like to be, and also looks through different channels to find out how she matches her own ideal.
So she’s reading a lot about what she can do, looking at what others she admires has done, and she’s decided she probably should up her skin care game. She’s reading about skin care routines and what you should do and what you should have. Ahh, she thinks she needs a serum.
she needs to find out about all the properties of these different skin care products to find out which one matches her needs. There are so many serums out there, she’s looking at specific ingredients and also looking at what the outcome of those serums is, does it create a result she was looking for? Is it going to feel better about taking proactive steps to take care of her skin?
So now she has done all her research and she’s found your serum, it meets all her criteria, makes her feel relieved about how well it works, and perhaps it’s a reputable and popular brand. For instance, using La Mer products is perceived differently from using Maybelline.
Notice that purchasing is always painful. And you could easily lose them at this step even if they had originally decided on buying your product
Now they are evaluating your product, does it make them feel better, are they among a certain enviable group of women that use it, are other people finding success as they did?
So bottom line, know what you want to know about your audience on a deeper level. Make sure that you are finding out about their lifestyle – go broader to know their lifestyle and their larger motivations in the field where your product or service is most applicable. Then segment them based on these different motivations or these different behaviours. And lastly, make sure that you are addressing the specific needs they have by understanding the purchase journey at a deeper level.
Now you know about who your audience is and what they want, fear, and hope for. Now what?
Now we can answer a bit about the person, our task is to figure out how we fit what we are offering into their lives to make it relevant. Is it a product, a service, even a message? Now we know what the context is so it’s a lot easier to start to position things based on what we know they will emotionally react to already.
We need to go back to their lifestyle as well as where you think they are on the purchasing journey. What information are they looking for? Does your product or service answer some of the questions they are asking currently? Does your product or offering occur in a related field that will generate talk and awareness? What emotions are your audience experiencing? Are we taking away a negative emotion and bringing relief? Or are we increasing someone’s positive feelings? We need to know the emotional reactions that are important to tap into for your audience.
A good example is Dove. They joined into the conversation and became relevant again because they weren’t talking about hygiene, which is what most of Dove’s products are for. What they did was align themselves with a lifestyle or an ideology that people can understand and identify with and enter into a conversation that people are already talking about, just not with brands. It’s a highly emotional topic, and hygiene is related to how people perceive each other and themselves. They just made hygiene meaningful on a very personal and empowering level and they did it without pushing their individual products or the benefits of those products. This is important in framing your content. What is the content that people will want to consume and how does your contribution add value to their life and incidentally, their perception of your brand in their life? Dove managed to create content that helped individuals align their brand with a lifestyle of authenticity and self-acceptance.
The other side of the coin is communicating with audiences at an implicit level. It’s not necessarily in what you say or how you say it. It might be in making an experience effortless and “intuitive.” People talk about intuition, and part of personalisation is not just in speaking their language, but providing people with things that they need when they need it. This is part of the big push now for personalisation – how do we do this? And many companies will swear that they’ve got the magic algorithm to make things perfectly personalised for each audience member. This is only half the story. One thing that I want to argue is that our AI is not intelligent enough to mimic human wants and needs. It’s getting better and better, but machines still have a very hard time understanding humans. They can’t perceive sarcasm, innuendo, and many of the subtleties of how humans communicate their feelings, wants, and desires. Our technology is extremely good at doing complex calculations quickly though. So if we give them the right inputs, they can output some amazing things to you, but we’re still at a point where we have to make sure that those inputs are highly meaningful.
One of the ways that this has been done is through taxonomies that are created by people. These taxonomies allow computers to find patterns that are as deep or as shallow as you want. It depends on the sophistication of the taxonomy. For instance, Pinterest uses folksonomies, where pictures are tagged but based on the community’s categorisation of the tags. So the more someone pins a photo to a specific board, the more knowledge is gained about that photo. For instance, even though I have a board called Style, because I’ve pinned photos that have been tagged by more specific board names, such as “minimalism” or “classic fashion”, over time, Pinterest starts to learn that even though I put that pin in “Style”, my “style” is actually an amalgamation of a specific type of style that others have categorised more specifically. They then know to serve me content that has similar tags – even if I didn’t ask for it and even though I am unaware of the tag it’s under.
And the mother of all deep level taxonomies is Netflix. According to netflix, 75% of its subscribers watch what is personally selected for them. This means that Netflix needs to be really, really good at recommending to keep people coming back. In fact, it may also be because of their amazing recommendation system that so many people trust and consume things that Netflix has recommended to them.
In 2006, Netflix had a contest with a reward of 1 million dollars to anyone who could create an algorithm that could accurately guess how someone would rate what they watched better than Netflix’s own algorithm. By 2007, over 20,000 teams from 150 countries had entered the competition. They awarded it in 2009 to BellKor for improving accuracy by 10% but by that time, Netflix had a system that was even more accurate. It couldn’t compete with their tagging system.
Their taxonomy system is highly sophisticated. They use over 1000 “microtags” where they tag almost everything you could imagine, from the plot to the likability of the main character. Because of that, they have created all these combinations together to form over 76,897 distinct movie categories. So Netflix can be super precise in the tastes that its viewers have.
So how does this tagging system work and why is it important in how we personalise content? So netflix tracked down a bunch of film aficionados that love to watch films and will pick out unusual details that we all might check out when we’re watching something, but don’t fall nicely into a “horror” genre. They tag deeply, with things like rating how romantic the movie is, what type of plot does it follow, and intelligent or cerebral the material is, and emotional things like the squirm factor of the situations the characters find themselves in. You’ve all seen those types of comedies, comedies where people are in absolutely embarrassing and humiliating situations.
Then they create genres based on all those taxonomies, these genres are highly detailed because the taxonomies are detailed.
Mother-Son Movies from the 1970s
Critically-acclaimed Emotional Underdog Movies
Gritty Suspenseful Revenge Westerns
Feel-good Romantic Spanish-Language TV Shows
Remember that all these tags and all these genres are behind the scenes. No one actually knows what the categories are, they just know that they are now getting the content they want.
Now they have the tools to understand their content on a very human level and their genres are so specific that they can provide a super accurate understanding of exactly the tastes and emotions that individuals prefer in the content they consume. Now all they need to know is what individuals are consuming and match those genres to the people. This is where technology shines. Technology can show how long they look at an item, what they search for, what they click on, what they stop watching, when they watch it. These are all important indicators of engagement with the material. Because they have the a deep knowledge of the content, they can piece together a deeper knowledge of the user based on these actions now. You can collect all the actions but not know what it means, but if you know what the content means on a personal level, then it makes sense.
Then you can display recommendations knowing what is driving them to watch certain things – the taxonomy helps in identifying underlying themes that might not be apparent in knowing their ratings of films.
And the crowning glory of Netflix is being able to use this rich knowledge to predict what people will want to watch, which is precisely what they did when they launched their first TV show – House of Cards.
"My first goal was: tear apart content!" – Todd Yellin, VP of product innovation, creator of the Netflix taxonomy
So what is the moral of this story? Technology is great – but personalisation still means adding in the human element. There is no algorithm thus far that understands the weird and quirky nature of how our brain works and processes things – not yet anyway. What we need to look into are hybrid solutions. This means providing a very well researched and thought out way of categorising information. So we need to understand firstly how people evaluate information (this is why you need to know your audience on a psychological level). Then you need to identify what behaviours are going to be important in understanding how people consume your content – is device important? Is time important? Once you’ve put those two things together, you’ve created something that can be automated, because it has human thinking behind it. So often, we create algorithms, and then try to figure out afterwards what those behaviours could have meant. We need to change that, by starting off with knowing our own content better.
A good example of this is with a client we had that would share inspirational online written or video stories. For instance, we could tag them based on surface level information like location, type of media used… but we can create deeper tags such as the emotions that are felt from these types of stories, the central character in the story, the writing style that is used. In fact, they use a type of folksonomy by allowing individuals to emotionally rate each article that they read. Again, these tags are in the CMS, they aren’t there necessarily for all the viewers to see. But they help our clients understand more about their users.
Here’s a good example of how we used categories to better understand the success of their content. For instance we worked with them to come up with a definition of a successful story. It had to be a story that people stayed on longer, was a unique page that they visited, and where people didn’t leave once they read it. Then we looked at how many words in the articles were considered humanizing. Meaning how many sentences in the whole article had words that related the reader to the main character of the story? We found that the more these types of sentences were used, the more likely people stayed on the story longer and didn’t leave once they finished reading. This is important in also identifying patterns in what people want when they consume content – on a much deeper level, not just based on location or word count.
This is the power of understanding your content deeply. It can help you understand new things that people are looking for. Many people think that data can only tell you what has already happened, and can’t predict more radical shifts. But taxonomy that is slightly more abstract, that taps into underlying motivations and fears can open up new possibilities. Motivations and fears don’t always go away with the advent of different products, they are usually constant drives that account for consumption behaviour of many things, old and new.
This was the case for Netflix, as I told you about their endorsement of House of Cards. In its first season, it received 9 Emmy nominations, winning 3. Moreover, it is the most popular series on all of Netflix currently. But they knew it would be a success even before they signed the papers because they understand psychologically what intrigues people and they even can guess how popular it will be based on the actors and director in it.
And the coup de grace for why you need to understand how people view content – the taxonomy was also helpful in targeting advertising of House of Cards so that trailers were more personalised based on what people liked to watch.
http://variety.com/2015/digital/news/netflix-originals-viewer-data-1201480234/
So the big bottom line – we need to go deeper in how we know our audience and how we know our content – this means looking into the psychology of our audiences, their loves, hates, dreams, and desires. We need to understand how they react emotionally, what scares them, what excites them, not JUST what it was they consumed. The way that we tailored messages should be embedded in a dialogue that is already important to your audience and this sometimes means it’s not about your product per se. And lastly, matching the depth of our content to the depth of our audience allows for much more relevant personalisation and future insight into what speaks to them most.
To sum up my talk, we need to look at the audience on a personal level.
We need to revisit how we do research and what questions are the right questions to ask. Do we ask questions on motivations, hurdles, and triggers? Do we know their self, emotional, and functional information needs?
Secondly, we need to understand that this is a process and the motivations, hurdles, triggers, and information needs change over time. Personalisation should help to identify those points to get it right.
And the second half of the equation is knowing your product as deeply as you should know your audience.
To personalise well, it needs to be a combination of both human understanding and machine strength. This means categorising content in ways similar to how your audience categorises them, deeply with psychological elements to them. It means knowing what the machine should be measuring.
Secondly, you need to revisit your content from the point of view of the audience member – why should they care about you? How do you fit in their lives and their conversations? And as netflix showed, you can do both, by using what you know from your rich, human-based data collection to understand what dialogue is going to ring true to its viewers.
SO the next time you’re faced with this challenge, remember to think about her motivations and emotions and how they might change over time.
Look at your product based specifically on how it meets her personal motivations and emotions.
Remember, personalisation means taking the pain out of searching and choosing. It IS about motivations and emotions - It should get us closer to our goals and make us feel happier about the decisions we make.