Advertisement
English transcript
DAVEY NICKELS: Please join me in welcoming Cole Nussbaumer
Knaflic, "Storytelling with Data" for Authors At.
[APPLAUSE]
Cole was on our team, people analytics and people operations
here at Google many, many years ago,
and she led our relationship with sales.
And she was known throughout the team
for her incredible expertise with presenting data,
visualizing data, and making it actually speak
to clients and to users.
And lots of funny stories about Cole, but one that stands out
is her feelings on pie charts.
I put together an analysis once that showed a distribution
of a specific population by level or something,
put it in a beautiful pie chart that I thought was amazing.
And Cole gave me very brutally honest feedback
about how ineffective it was in actually showing
the differences in the group.
So Cole, I'm very curious to hear today
if you still have those views on pie charts,
or if maybe they're a little bit more acceptable now.
COLE NUSSBAUMER KNAFLIC: I still have a pretty strong view
when it comes to pie charts.
DAVEY NICKELS: OK.
Well, we look forward to hearing a little more about that.
So the way that this is going to work
is we're going to have 30 minutes of a lesson, actually,
from Cole from her book.
And then second, we're going to have Q&A, moderated
by Tina Malm, from people analytics and me,
Davey Nickels, from people analytics as well.
So please, again, join me in welcoming Cole
to her first stop on her book tour here at Google.
[APPLAUSE]
COLE NUSSBAUMER KNAFLIC: Awesome.
So it seems so fitting for me for Google
to be the first stop after the official publication
of my book, since this is where so much of it began.
I started at Google back in 2007 on the people analytics team.
And people analytics is an analytical team
in the people ops organization here,
where the goal is to try to ensure
that all the decisions about people--
employees, future employees-- are data-driven.
Now, I had the opportunity of joining
when the team was relatively small,
which meant that I got to work on a ton of cool stuff
over the years, learning about things like what
makes a manager effective, how do you build productive teams,
what drives attrition.
Now, in 2010, we developed a program called Base Camp.
It was an internal MBA-like training program
within people operations.
And I was asked to build content on data visualization, which
was an awesome opportunity, because I'd always been really
interested in this space.
But this meant I could pause and research and understand
why some of the things I'd arrived
at through trial and error over time had been effective.
So as a side note, I can honestly say that Google for me
was life-changing.
I didn't know it at the time, but the very first time
I delivered the Data Visualization course, which
was at a people ops off-site in Monterrey, in the very first
row sat my future husband.
So you can really say without the Data Visualization course
here at Google, I wouldn't have these
or the third that's on the way-- super life-changing.
But back on Google, there was broader interest,
we were finding, in the Data Visualization course.
So we actually ended up rolling it out
across Google, which meant I got the opportunity
to travel to offices in London and Dublin and Zurich and train
trainers and teach courses.
I also got a chance to teach courses
across a number of US offices, including many right
here in Mountain View.
One of the things that was interesting for me
was to see salespeople and engineers sitting side
by side in those courses.
I came to realize that the skills needed in this area
were fundamental.
They weren't specific to any given role.
They also weren't specific to Google.
Over time, other organizations started reaching out to me,
wanting me to go teach their teams and their organizations
how to communicate effectively with data.
So over the course of the past couple of years,
I've had the opportunity to work with hundreds
of different teams across many different organizations.
And usually this takes the form of workshops,
where I'll spend half a day or a full day teaching
foundational lessons on communicating with data.
And oftentimes what I'll do is solicit examples
from the group ahead of time, and we'll
go through the lessons, and then we'll
spend time talking about that group's specific examples.
And I'll go through my makeovers as one potential path
that leverages the lessons that we've covered.
So I thought it might be cool to take you
through a couple of these.
We won't go through them in a lot of detail.
We'll just give you a visual sense of what
you can learn in the book.
So this first example is one from the philanthropic sector.
So it's a foundation that wanted to start a conversation
on shifting spending from non-initiative, which
is the big, cream-colored segment at the top left,
to higher education, that tiny, blue segment.
So in this case, we change from the pie chart--
we've already talked about my views there-- to this, right?
If we want to shift spending, let's
say we want to shift spending and start
a conversation about that.
Use visual cues to draw our audience's attention to where
we want them to pay it.
Let's look another example.
This is one from an IT group who couldn't believe that people,
after looking at this graph, weren't doing anything
with this information.
Why weren't they acting upon this information?
Here, it gets totally lost in the graph, right?
There's no story to bring it to life.
Now, the backstory was-- if we stare at this long enough,
we can see there's a gap starting to form out
on the right-hand side, where the number of tickets coming in
are exceeding the number that are being processed.
Now, the backstory was they'd lost a couple of resources.
They were understaffed.
And they really wanted to hire a couple more people.
And they couldn't understand why nobody was doing that
based on this graph.
So in this case, we changed from that to this,
again making that call to action clear,
annotating the context directly on the graph,
drawing attention to that gap that's forming
on the right-hand side.
So I've got one more of these.
This is an example from a small organization who
was recognizing that their regional sale composition had
shifted over time and wanting to have
a conversation about what some of the implications of that
were.
Now, this is an interesting one.
I've used this a number of times in workshops.
And when people flip to this graph,
there's often a sort of immediate, negative, visceral
reaction, which is something we want to try
to avoid in our audiences.
And now, I can't imagine all of them
were Packers fans like my husband.
These are sort of Broncos colors, right?
But rather that this graph was unnecessarily confusing.
Now, the beautiful thing here is there are some clear take-aways
articulated at the bottom.
It's just almost impossible to know
where to look in the data for evidence of those take-aways.
So in this case, changed from that to this,
making the focus on the change, tying the text
directly to the data, both through proximity
and similarity of color.
Now, one thing to note is these examples
cross many different industries.
I mentioned before that these skills are not role-specific.
They're also not really industry-specific.
Rather, they're foundational.
And over time, through all of the workshops
that I've taught over the past few years,
I've codified these lessons.
And that's what ultimately led to this, my book.
And I'm super excited to be able to share with you today
a couple quick lessons from my new book,
"Storytelling with Data."
So we're going to talk today about two key lessons,
first off, focusing attention, and secondly, telling a story.
We want to draw one important distinction
at this point, which is the distinction
between exploratory analysis and explanatory analysis.
So exploratory analysis, you perhaps start off
with a question or hypothesis, or you're just
digging through your data, trying
to understand what's interesting,
what can I learn about this data that somebody else
might care about?
Once you've identified that interesting thing,
then we move into explanatory space.
That is where you have something specific
you want to communicate to somebody specific.
And it's this latter space we want to keep in mind today.
And when it comes to explanatory analysis,
these lessons become more important than perhaps
any others-- first off, thinking about where
you want your audience to pay attention
and doing things on purpose to make that happen,
and then secondly, never simply showing data, but rather
making data a pivotal point in an overarching story.
So we'll talk briefly through each of these.
First off, focusing attention-- I
can recall a time at Google where I was working
on the Project Oxygen study.
Quick show of hands-- how many people
are familiar with Project Oxygen?
Most people in the room.
So Project Oxygen was a study led by my colleague Neil Patel.
And the goal really was to try to understand,
on a mathematical, statistical level,
what makes managers effective.
One of the challenges that we encountered
was, after the study was done, communicating the results of it
to a very mixed audience, where we had both engineers, who
had a great desire for detail-- they wanted to understand
the methodology, they wanted to be
convinced of the robustness of the analysis.
At the same time, we were also wanting
to communicate to sales managers,
for example, who were mostly less concerned with methodology
and more concerned about what's in it for me.
How should I act differently based on this?
And so what we found was by really being careful about how
we focused attention, we could preserve a lot of that detail
but push it to the background and make the meta-point pop
out so that it was clear.
Let's talk a little bit about how
people see to get into more of this how we focus attention.
So here's a super-simplified picture of that process.
On the left-hand side, you have light refracting off
a stimulus.
This gets captured by our eyes.
We don't fully see with our eyes.
Rather, most of what we think of as visual processing
takes place in our brains.
Now, in the brains there are a few types of memory
that are important to understand as we're designing
visual communications.
We'll talk about one of them today, which is iconic memory.
Iconic memory is super short-term.
It's shorter than short-term memory,
and information stays there for fractions of a second
before it gets forwarded on to our short-term memory.
The really cool thing about iconic memory
is that it's tuned to a specific set of what we
call pre-attentive attributes.
So let's actually pause here and do a quick exercise.
So in a moment, I'm going to put a bunch of numbers
up on the screen.
What I'd like you all to do as fast as you can
is count the number of 3s that you see.
We got it?
We're going to count 3s.
When you know the answer, shout it out.
It is a race.
You would like to win.
Ready, set, go!
All right, six is the correct answer.
This took a bit of time, though, right?
You had to physically read through these four
lines of text, look for a 3, which
is kind of a complicated shape.
Watch what a different exercise it
becomes when I make one tiny change.
Don't have time to think.
Don't have time to blink.
Suddenly there are six 3s in front of you.
This is so apparent so quickly because I'm
leveraging your iconic memory.
I'm using the pre-attentive attribute
of intensity of color, in this case,
to make the 3s the one thing that stand out as
different from the rest.
Now, this is hugely critical, because what this means is
our pre-attentive attributes, if we use them strategically,
can help enable us to get our audience
to see what we want them to see before they even
know they're seeing it.
Here are the attributes.
I won't read through all of these,
but notice as your eyes scan across the screen
how they're just drawn to the one
within each group that's different from the rest.
You don't really have to consciously look for it.
Now, one thing to know about the attributes
is people tend to associate quantitative values
with some but not others.
For example, most people will consider a long line
to represent a greater value than a short line.
It's one of the reasons bar charts are
intuitive for us to read.
But we don't think of hue, for example, in the same way.
If I ask you which is greater, red or blue,
it's not really a meaningful question.
And this is important because it tells us
which of the attributes can be used
in code quantitative information and which should be used
as categorical differentiators.
Now as you can perhaps imagine, pre-attentive attributes
become huge tools for focusing our audience's attention when
it comes to visualizing data.
So here's some sort of generic data from our annual customer
survey.
We can see how we've fared across a number of dimensions.
Notice how without other visual cues, this becomes very much
like the count-the-3s example again.
We have to look at this data, read through it,
figure out what might be important to pay attention to.
Whereas if I'm the one communicating this data,
I should have already done that for you, in which case
I can use some pre-attentive attributes, perhaps paired
with some explanatory text, to draw your attention very
quickly to one part of the story, right?
Price and convenience-- we're doing awesome here.
Let's pause and celebrate our success.
Or I can use this same broad strategy
to draw your attention to a totally different place
in the data.
But we're struggling when it comes
to relationship and brand.
How can we positively impact these areas?
Now, there's a test I like to employ in trying to figure out
whether you're using your pre-attentive attributes
strategically.
And that is the Where Are Your Eyes Drawn Test,
where you look away from your visual and look back at it,
or close your eyes and look back at it,
and just notice where your eyes land first.
Because it's probably where your audience's eyes
will land as well.
So I thought we'd do this with a series of pictures
and just talk about the implications
for our visual designs.
So I'm going to put a series of different pictures up there.
When I put the picture up, just shout out
where your eyes go first.
Ready?
Where do your eyes go first here?
AUDIENCE: Stop.
COLE NUSSBAUMER KNAFLIC: Stop sign, right?
You almost can't look anywhere else at the onset,
because it's bright, it's red.
It's got these big, bold, capital letters.
It's outlined in white, which sets it apart
from the background.
We want to think about how you can use some of those cues
when you're visualizing data to draw attention as well.
Let's do another one.
Where your eyes go here?
AUDIENCE: The sun.
COLE NUSSBAUMER KNAFLIC: Yeah, if you're like most people,
they go to the sun.
But if you're like me, when you're
trying to look at the sun, you get this plane sort
of tugging on your peripheral vision.
Or if I try to look at the plane,
I can see the sun sort of wanting to pull me that way.
So just be aware that when you're
emphasizing multiple things in a graph or on a page,
this tension can be created in your audience.
How about where do your eyes go here?
It depends a little bit, perhaps,
on where you're sitting in the room.
A lot of people will be drawn to that Perennial Sale
sign in bright pink, because it's bright,
because of the black, bold lettering on it.
And then most people from there will
continue down and rightward.
And that's because, without other visual cues,
we typically start at the top left of our page or our screen
and do zigzagging Zs across.
So in this case, that draw to the Perennial Sale was strong.
We started there and then continued on that Z downward
and to the right.
Notice that means we missed whatever
was happening in the top left quadrant, and maybe
that second and third quadrant as well.
So to be thoughtful about the overarching designs
of the pages on which your data visualization sits
and take that into account.
Just a couple more of these-- where do your eyes go here?
Everywhere and nowhere all at the same time.
Colorful is an awesome goal for a birthday party.
Color is not such an awesome goal
when it comes to visualizing data.
When we make so many things different,
we have a lot of stuff competing for our attention, which
actually makes it really hard to look at any one thing.
Check out the difference in how your attention is focused here
versus here, right?
With the red balloon, the one thing
that's different on the whole page
we almost can't not look at it.
That is the power of color specifically,
used strategically.
Let's take a look now at an example
from that Project Oxygen study that I mentioned at the onset.
This is what one of our original slides looked like.
It's been genericized a bit.
We can see our main takeaway at the top--
some elements of job satisfaction
are more sensitive to manager quality than others.
We've got some categorizations here and then our data
at the bottom.
Here we're not using color so strategically.
Here, color is used as a categorical differentiator.
There originally, we've taken them off here
but were categories along the bottom.
You can think of those like Googlegeist categories,
things like career development and performance management
and culture, which is not necessarily how
we want to be using our color.
So in this case, our redesign looked like this.
The graph is mostly the same.
The contents of the page are pretty much exactly the same.
We've just rearranged things a bit
and used our pre-attentive attributes, color specifically,
more thoughtfully to really draw our audience's attention
to where we want them to pay it.
While we draw attention, we also want
to think about embedding our data in story.
By way of a Google anecdote, I can recall a time
when I was working with one of the junior analysts
on our team.
And she had just finished analyzing Googlegeist results,
results from the annual employee survey [INAUDIBLE] part
of the organization and was getting
ready to communicate those results
to the leader of that team.
And this particular team had been
struggling in a lot of places.
The scores weren't great.
So there was some sensitivity around how that message should
be delivered.
And the deck at that point was page
after page after page of the standard report--
no story and little narrative to tie it all together.
It would have been very easy for the leader of that group
to say, well, that's interesting,
and move on to the next thing.
That would have been a failure.
So what I had the analyst do was set the deck aside and tell me
the story.
Tell me what you learned when you were analyzing this data.
And when we did that, the articulation of the story
was super powerful.
There were clear areas for improvement,
and she knew exactly where to focus action
to achieve that improvement.
This we could use to light a fire
under the leader for that team.
So it's a good example of how data without story
isn't always so meaningful.
But the story can help bring the data to life.
So we want to think about how we can leverage that power when
we are communicating with data every time we're doing it.
Here are some facts on a slide.
Go ahead and read through these.
Anybody recognize what we're looking at here?
What story is this?
Red Riding Hood, right?
But facts on a slide are not so compelling or memorable.
If I ask you a day or two from now,
what distance was it from Red Riding Hood's
house to grandma's, or what time did Red Riding Hood get there,
these aren't likely facts that you will
have committed to your memory.
Stories, on the other hand, are memorable.
How many people-- quick show of hands--
know the story of Red Riding Hood?
Pretty much everybody in the room.
We'll do a quick thought exercise here.
We'll just take about 15 seconds.
Close your eyes, or stare up at this screen,
and I'd like you to recall for yourself the story of Reg
Riding Hood, thinking specifically
about the plot,
…
Knaflic, "Storytelling with Data" for Authors At.
[APPLAUSE]
Cole was on our team, people analytics and people operations
here at Google many, many years ago,
and she led our relationship with sales.
And she was known throughout the team
for her incredible expertise with presenting data,
visualizing data, and making it actually speak
to clients and to users.
And lots of funny stories about Cole, but one that stands out
is her feelings on pie charts.
I put together an analysis once that showed a distribution
of a specific population by level or something,
put it in a beautiful pie chart that I thought was amazing.
And Cole gave me very brutally honest feedback
about how ineffective it was in actually showing
the differences in the group.
So Cole, I'm very curious to hear today
if you still have those views on pie charts,
or if maybe they're a little bit more acceptable now.
COLE NUSSBAUMER KNAFLIC: I still have a pretty strong view
when it comes to pie charts.
DAVEY NICKELS: OK.
Well, we look forward to hearing a little more about that.
So the way that this is going to work
is we're going to have 30 minutes of a lesson, actually,
from Cole from her book.
And then second, we're going to have Q&A, moderated
by Tina Malm, from people analytics and me,
Davey Nickels, from people analytics as well.
So please, again, join me in welcoming Cole
to her first stop on her book tour here at Google.
[APPLAUSE]
COLE NUSSBAUMER KNAFLIC: Awesome.
So it seems so fitting for me for Google
to be the first stop after the official publication
of my book, since this is where so much of it began.
I started at Google back in 2007 on the people analytics team.
And people analytics is an analytical team
in the people ops organization here,
where the goal is to try to ensure
that all the decisions about people--
employees, future employees-- are data-driven.
Now, I had the opportunity of joining
when the team was relatively small,
which meant that I got to work on a ton of cool stuff
over the years, learning about things like what
makes a manager effective, how do you build productive teams,
what drives attrition.
Now, in 2010, we developed a program called Base Camp.
It was an internal MBA-like training program
within people operations.
And I was asked to build content on data visualization, which
was an awesome opportunity, because I'd always been really
interested in this space.
But this meant I could pause and research and understand
why some of the things I'd arrived
at through trial and error over time had been effective.
So as a side note, I can honestly say that Google for me
was life-changing.
I didn't know it at the time, but the very first time
I delivered the Data Visualization course, which
was at a people ops off-site in Monterrey, in the very first
row sat my future husband.
So you can really say without the Data Visualization course
here at Google, I wouldn't have these
or the third that's on the way-- super life-changing.
But back on Google, there was broader interest,
we were finding, in the Data Visualization course.
So we actually ended up rolling it out
across Google, which meant I got the opportunity
to travel to offices in London and Dublin and Zurich and train
trainers and teach courses.
I also got a chance to teach courses
across a number of US offices, including many right
here in Mountain View.
One of the things that was interesting for me
was to see salespeople and engineers sitting side
by side in those courses.
I came to realize that the skills needed in this area
were fundamental.
They weren't specific to any given role.
They also weren't specific to Google.
Over time, other organizations started reaching out to me,
wanting me to go teach their teams and their organizations
how to communicate effectively with data.
So over the course of the past couple of years,
I've had the opportunity to work with hundreds
of different teams across many different organizations.
And usually this takes the form of workshops,
where I'll spend half a day or a full day teaching
foundational lessons on communicating with data.
And oftentimes what I'll do is solicit examples
from the group ahead of time, and we'll
go through the lessons, and then we'll
spend time talking about that group's specific examples.
And I'll go through my makeovers as one potential path
that leverages the lessons that we've covered.
So I thought it might be cool to take you
through a couple of these.
We won't go through them in a lot of detail.
We'll just give you a visual sense of what
you can learn in the book.
So this first example is one from the philanthropic sector.
So it's a foundation that wanted to start a conversation
on shifting spending from non-initiative, which
is the big, cream-colored segment at the top left,
to higher education, that tiny, blue segment.
So in this case, we change from the pie chart--
we've already talked about my views there-- to this, right?
If we want to shift spending, let's
say we want to shift spending and start
a conversation about that.
Use visual cues to draw our audience's attention to where
we want them to pay it.
Let's look another example.
This is one from an IT group who couldn't believe that people,
after looking at this graph, weren't doing anything
with this information.
Why weren't they acting upon this information?
Here, it gets totally lost in the graph, right?
There's no story to bring it to life.
Now, the backstory was-- if we stare at this long enough,
we can see there's a gap starting to form out
on the right-hand side, where the number of tickets coming in
are exceeding the number that are being processed.
Now, the backstory was they'd lost a couple of resources.
They were understaffed.
And they really wanted to hire a couple more people.
And they couldn't understand why nobody was doing that
based on this graph.
So in this case, we changed from that to this,
again making that call to action clear,
annotating the context directly on the graph,
drawing attention to that gap that's forming
on the right-hand side.
So I've got one more of these.
This is an example from a small organization who
was recognizing that their regional sale composition had
shifted over time and wanting to have
a conversation about what some of the implications of that
were.
Now, this is an interesting one.
I've used this a number of times in workshops.
And when people flip to this graph,
there's often a sort of immediate, negative, visceral
reaction, which is something we want to try
to avoid in our audiences.
And now, I can't imagine all of them
were Packers fans like my husband.
These are sort of Broncos colors, right?
But rather that this graph was unnecessarily confusing.
Now, the beautiful thing here is there are some clear take-aways
articulated at the bottom.
It's just almost impossible to know
where to look in the data for evidence of those take-aways.
So in this case, changed from that to this,
making the focus on the change, tying the text
directly to the data, both through proximity
and similarity of color.
Now, one thing to note is these examples
cross many different industries.
I mentioned before that these skills are not role-specific.
They're also not really industry-specific.
Rather, they're foundational.
And over time, through all of the workshops
that I've taught over the past few years,
I've codified these lessons.
And that's what ultimately led to this, my book.
And I'm super excited to be able to share with you today
a couple quick lessons from my new book,
"Storytelling with Data."
So we're going to talk today about two key lessons,
first off, focusing attention, and secondly, telling a story.
We want to draw one important distinction
at this point, which is the distinction
between exploratory analysis and explanatory analysis.
So exploratory analysis, you perhaps start off
with a question or hypothesis, or you're just
digging through your data, trying
to understand what's interesting,
what can I learn about this data that somebody else
might care about?
Once you've identified that interesting thing,
then we move into explanatory space.
That is where you have something specific
you want to communicate to somebody specific.
And it's this latter space we want to keep in mind today.
And when it comes to explanatory analysis,
these lessons become more important than perhaps
any others-- first off, thinking about where
you want your audience to pay attention
and doing things on purpose to make that happen,
and then secondly, never simply showing data, but rather
making data a pivotal point in an overarching story.
So we'll talk briefly through each of these.
First off, focusing attention-- I
can recall a time at Google where I was working
on the Project Oxygen study.
Quick show of hands-- how many people
are familiar with Project Oxygen?
Most people in the room.
So Project Oxygen was a study led by my colleague Neil Patel.
And the goal really was to try to understand,
on a mathematical, statistical level,
what makes managers effective.
One of the challenges that we encountered
was, after the study was done, communicating the results of it
to a very mixed audience, where we had both engineers, who
had a great desire for detail-- they wanted to understand
the methodology, they wanted to be
convinced of the robustness of the analysis.
At the same time, we were also wanting
to communicate to sales managers,
for example, who were mostly less concerned with methodology
and more concerned about what's in it for me.
How should I act differently based on this?
And so what we found was by really being careful about how
we focused attention, we could preserve a lot of that detail
but push it to the background and make the meta-point pop
out so that it was clear.
Let's talk a little bit about how
people see to get into more of this how we focus attention.
So here's a super-simplified picture of that process.
On the left-hand side, you have light refracting off
a stimulus.
This gets captured by our eyes.
We don't fully see with our eyes.
Rather, most of what we think of as visual processing
takes place in our brains.
Now, in the brains there are a few types of memory
that are important to understand as we're designing
visual communications.
We'll talk about one of them today, which is iconic memory.
Iconic memory is super short-term.
It's shorter than short-term memory,
and information stays there for fractions of a second
before it gets forwarded on to our short-term memory.
The really cool thing about iconic memory
is that it's tuned to a specific set of what we
call pre-attentive attributes.
So let's actually pause here and do a quick exercise.
So in a moment, I'm going to put a bunch of numbers
up on the screen.
What I'd like you all to do as fast as you can
is count the number of 3s that you see.
We got it?
We're going to count 3s.
When you know the answer, shout it out.
It is a race.
You would like to win.
Ready, set, go!
All right, six is the correct answer.
This took a bit of time, though, right?
You had to physically read through these four
lines of text, look for a 3, which
is kind of a complicated shape.
Watch what a different exercise it
becomes when I make one tiny change.
Don't have time to think.
Don't have time to blink.
Suddenly there are six 3s in front of you.
This is so apparent so quickly because I'm
leveraging your iconic memory.
I'm using the pre-attentive attribute
of intensity of color, in this case,
to make the 3s the one thing that stand out as
different from the rest.
Now, this is hugely critical, because what this means is
our pre-attentive attributes, if we use them strategically,
can help enable us to get our audience
to see what we want them to see before they even
know they're seeing it.
Here are the attributes.
I won't read through all of these,
but notice as your eyes scan across the screen
how they're just drawn to the one
within each group that's different from the rest.
You don't really have to consciously look for it.
Now, one thing to know about the attributes
is people tend to associate quantitative values
with some but not others.
For example, most people will consider a long line
to represent a greater value than a short line.
It's one of the reasons bar charts are
intuitive for us to read.
But we don't think of hue, for example, in the same way.
If I ask you which is greater, red or blue,
it's not really a meaningful question.
And this is important because it tells us
which of the attributes can be used
in code quantitative information and which should be used
as categorical differentiators.
Now as you can perhaps imagine, pre-attentive attributes
become huge tools for focusing our audience's attention when
it comes to visualizing data.
So here's some sort of generic data from our annual customer
survey.
We can see how we've fared across a number of dimensions.
Notice how without other visual cues, this becomes very much
like the count-the-3s example again.
We have to look at this data, read through it,
figure out what might be important to pay attention to.
Whereas if I'm the one communicating this data,
I should have already done that for you, in which case
I can use some pre-attentive attributes, perhaps paired
with some explanatory text, to draw your attention very
quickly to one part of the story, right?
Price and convenience-- we're doing awesome here.
Let's pause and celebrate our success.
Or I can use this same broad strategy
to draw your attention to a totally different place
in the data.
But we're struggling when it comes
to relationship and brand.
How can we positively impact these areas?
Now, there's a test I like to employ in trying to figure out
whether you're using your pre-attentive attributes
strategically.
And that is the Where Are Your Eyes Drawn Test,
where you look away from your visual and look back at it,
or close your eyes and look back at it,
and just notice where your eyes land first.
Because it's probably where your audience's eyes
will land as well.
So I thought we'd do this with a series of pictures
and just talk about the implications
for our visual designs.
So I'm going to put a series of different pictures up there.
When I put the picture up, just shout out
where your eyes go first.
Ready?
Where do your eyes go first here?
AUDIENCE: Stop.
COLE NUSSBAUMER KNAFLIC: Stop sign, right?
You almost can't look anywhere else at the onset,
because it's bright, it's red.
It's got these big, bold, capital letters.
It's outlined in white, which sets it apart
from the background.
We want to think about how you can use some of those cues
when you're visualizing data to draw attention as well.
Let's do another one.
Where your eyes go here?
AUDIENCE: The sun.
COLE NUSSBAUMER KNAFLIC: Yeah, if you're like most people,
they go to the sun.
But if you're like me, when you're
trying to look at the sun, you get this plane sort
of tugging on your peripheral vision.
Or if I try to look at the plane,
I can see the sun sort of wanting to pull me that way.
So just be aware that when you're
emphasizing multiple things in a graph or on a page,
this tension can be created in your audience.
How about where do your eyes go here?
It depends a little bit, perhaps,
on where you're sitting in the room.
A lot of people will be drawn to that Perennial Sale
sign in bright pink, because it's bright,
because of the black, bold lettering on it.
And then most people from there will
continue down and rightward.
And that's because, without other visual cues,
we typically start at the top left of our page or our screen
and do zigzagging Zs across.
So in this case, that draw to the Perennial Sale was strong.
We started there and then continued on that Z downward
and to the right.
Notice that means we missed whatever
was happening in the top left quadrant, and maybe
that second and third quadrant as well.
So to be thoughtful about the overarching designs
of the pages on which your data visualization sits
and take that into account.
Just a couple more of these-- where do your eyes go here?
Everywhere and nowhere all at the same time.
Colorful is an awesome goal for a birthday party.
Color is not such an awesome goal
when it comes to visualizing data.
When we make so many things different,
we have a lot of stuff competing for our attention, which
actually makes it really hard to look at any one thing.
Check out the difference in how your attention is focused here
versus here, right?
With the red balloon, the one thing
that's different on the whole page
we almost can't not look at it.
That is the power of color specifically,
used strategically.
Let's take a look now at an example
from that Project Oxygen study that I mentioned at the onset.
This is what one of our original slides looked like.
It's been genericized a bit.
We can see our main takeaway at the top--
some elements of job satisfaction
are more sensitive to manager quality than others.
We've got some categorizations here and then our data
at the bottom.
Here we're not using color so strategically.
Here, color is used as a categorical differentiator.
There originally, we've taken them off here
but were categories along the bottom.
You can think of those like Googlegeist categories,
things like career development and performance management
and culture, which is not necessarily how
we want to be using our color.
So in this case, our redesign looked like this.
The graph is mostly the same.
The contents of the page are pretty much exactly the same.
We've just rearranged things a bit
and used our pre-attentive attributes, color specifically,
more thoughtfully to really draw our audience's attention
to where we want them to pay it.
While we draw attention, we also want
to think about embedding our data in story.
By way of a Google anecdote, I can recall a time
when I was working with one of the junior analysts
on our team.
And she had just finished analyzing Googlegeist results,
results from the annual employee survey [INAUDIBLE] part
of the organization and was getting
ready to communicate those results
to the leader of that team.
And this particular team had been
struggling in a lot of places.
The scores weren't great.
So there was some sensitivity around how that message should
be delivered.
And the deck at that point was page
after page after page of the standard report--
no story and little narrative to tie it all together.
It would have been very easy for the leader of that group
to say, well, that's interesting,
and move on to the next thing.
That would have been a failure.
So what I had the analyst do was set the deck aside and tell me
the story.
Tell me what you learned when you were analyzing this data.
And when we did that, the articulation of the story
was super powerful.
There were clear areas for improvement,
and she knew exactly where to focus action
to achieve that improvement.
This we could use to light a fire
under the leader for that team.
So it's a good example of how data without story
isn't always so meaningful.
But the story can help bring the data to life.
So we want to think about how we can leverage that power when
we are communicating with data every time we're doing it.
Here are some facts on a slide.
Go ahead and read through these.
Anybody recognize what we're looking at here?
What story is this?
Red Riding Hood, right?
But facts on a slide are not so compelling or memorable.
If I ask you a day or two from now,
what distance was it from Red Riding Hood's
house to grandma's, or what time did Red Riding Hood get there,
these aren't likely facts that you will
have committed to your memory.
Stories, on the other hand, are memorable.
How many people-- quick show of hands--
know the story of Red Riding Hood?
Pretty much everybody in the room.
We'll do a quick thought exercise here.
We'll just take about 15 seconds.
Close your eyes, or stare up at this screen,
and I'd like you to recall for yourself the story of Reg
Riding Hood, thinking specifically
about the plot,
…