Ten years, and the 944,357 words, are proof that I love purposeful data, collecting it, pouring smart strategies into analyzing it, and using the insights identified to transform organizations.
In the quest for that last important bit, I am insanely obsessive about 1. simplification and 2. pressing the right emotional buttons.
The reasons are that we all like complexity, it gives us energy :), we tend to be logical, and we often treat data output as the end when in reality the data output is just the start of the process that results in actions that deliver business impact.
Very often the output of our work with Big Data or Small Data, Google Analytics or R, will end up in a few cells of a spreadsheet or a table in Word/Keynote/PowerPoint. The stakes for this output are higher when we are in front of the Senior Leadership of any company, we have but a few minutes to communicate what we have to. Hence my two obsessions above.
In this post, with lots of pictures and real-world data examples, I want to share 6 different strategies you can leverage in service of simplification and pressing the right emotional buttons. Along our journey, I've also sprinkled in 15 universal truths that will bring you joy.
Here are the sections in this post:
Death at the last-mile.
1. Rebel against crapification via cluttering.
2. Don't fragment data, don't forget higher order bits.
3. Obsess with deleting information provided.
4. Don't run away, make the tough choices.
5. So what? So What?? So WHAT!
6. Shift your worldview to your client's.
But Wait, There's More!
I've written a couple other posts in this same spirit, with 17 additional tips. I'll link to those posts and an article in Forbes in the last section.
You'll also find two examples below where I want you to play along. I'll share the before version, and I want you to challenge yourself, create a better after version, and share it with me. I'll add a couple of the best in each case below with your name (and your permission).
[update]
Clearly the readers of Occam's Razor are engaged, I got an avalanche of submissions! Thank you. I've added a selection of eleven below to inspire you. There are so many different, and simpler, way to solve the problem. In each case below, you'll see a section called update, as is the case here, and that's where you'll find the wonderful submissions from our readers along with a after version I'd originally created.
[/update]
Before we head any further I should stress that all the examples below are real. Some are from slides I saw at conferences, some from our company, and some from yours. In all cases, I've spent time changing company names or the names of the things on the slide (so if you see Display Ads, that might be TV data). I did not want to call anyone out, we all suck at this and we all need to get better together.
I'm assuming for the rest of this post that either you are the person responsible for presenting the output of the analytical efforts OR you are working with a person who will present and that they are curious and will commit the time and intellectual horsepower required.
99% of the final data outputs are heartbreakingly bad because the person telling the story did not create the story (thus the intellectual vacuum) and they furthermore did not commit the time required to understand the story enough to make it their own, to be able to tell it with passion/commitment.
The end result in all these cases is that data efforts come to naught. It is a simple yet hard to solve problem. Commit to addressing it, or at least exposing it and crap will not exist.
For the purpose of this post, I'm going to assume it does not exist.
Let's go, and have some sexy data presentation fun!
This is what I mean by an Analyst's death at the last mile… I'm confident a ton of work went into getting the data, analyzing it, creating this graph and choosing just the right fonts…
It is also true that it is completely ineffective and no one will understand anything.
Too many bars, inside them too many slices, odd color choices, all end up with this question: what the heck's going on here?
This whole thing should probably be just three bars (or if you really push me, four). In its current form it will take too long to explain and as a result you are going to lose the audience pretty quickly. And, this is likely 1 of 45 slides you have.
That is what I mean bydeath at the last-mile . A lot of great effort went into this, with no impact.
Or, checkout this beauty…
It breaks my heart. The Analyst who was presenting this at the conference was really smart. The output above is trying to do too many things, there are too many things on it, and the outcome will be that your audience of 5 or 5000 is on their phone reconsidering their dating pool.
The above should be a table with two rows and three columns. That's it. Obsessively simple, pressing the right emotional buttons.
One more.
Just because this one is from one of the largest business consulting companies in the world, and the title of the study was: Cutting Complexity and Adding Value. #ironic
There is only one simple message above, and just two metrics that matter. It should have just focused on those two. Everything else is a distraction or a ill-conceived strategy to prove the hourly consulting rate is worth it by pummeling the audience by taking up all available space on the slide.
In each case of death by last mile, how did we end up quickly identifying that only two numbers are important above or that the Cancer one has three columns or first one is best as three bars?
Rules learned from a lifetime of practice. And, an unhealthy obsession with simplicity.
Here are some of them, more at the end of this post.
1. Rebel against crapification via cluttering.
This is such a simple rule, and one that will fix any problem. Just don't clutter. Don't. Just don't.
Negative space is a good thing. Don't feel obligated to use every pixel available.
Know the difference between a handout – to be printed, to be shared after, to be sent off to lots of people who will not read it – and a presentation.
Universal Truth #1: You are the presentation, not the slides.
100% of the most memorable things that will be retained by your audience are going to come out of your mouth and won't be on the thing in the screen. I know. Even with the greatest data in the universe on the screen. You are the presentation. Please keep that in mind as we go and look at how to do the slide part, the table part, the analysis part smarter.
Let's look at this example, the research done is expensive and good. A copious amount of effort was put in by the Research Analysts. The whole thing dies at the last mile…
Universal Truth #2: Clipart is for adult earthlings with questionable taste (or young earthlings who don't know any better).
Universal Truth #3: Let things upset you. It is how you make things better.
Setting aside universal truth #2, it irritated me that you can't actually hold your phone like the lady in the clipart. Take out your phone. Try it. Position your fingers exactly like her. Are you feeling weird? Where's your thumb? See? Now, are you irritated like I am? Be irritated. Take that irritation, fix things. It is the path to glory.
Your goal is to present the data, as simply as possible. Let's KonMari!
Universal Truth #4: Don't puke all the data out at one.
Animations are your best friends. Control the pace of the story, let things be reveled as you speak/present, keep the focus on you.
Universal Truth #5: The only acceptable Animation on a slide is Fade.
The first thing we do is declutter the title (if you insist on having a title, I dislike them quite a bit as demonstrated by all the after examples in this post). Here it is…
The simplest way to communicate effectively is to have a structure that people can internalize and follow easily. In this case, with just a couple of minutes of thought we can come up with a simple structure to communicate the findings.
Share the first one, Shops…
Now dwell on it. There is nothing else to read on the slide. Just one insight. Talk about why it is interesting. Talk about what the audience can do with it. Talk about what shifts it represents. So on and so forth.
The process is how you'll establish your personal value. Nothing in the presentation, on the slide, no matter how compelling, will do that. The context you surround the data with, your words, will do that.
I won't stretch the entire process, but you'll present one element of the structure at a time, and this is where your Ashley slide will end up…
Shops, Loves, Watching.
Three relatable elements of a structure, they are sticky, and an insight under each.
Scroll back up. Look at the original.
Would you have gotten this simple and direct message from the clipart and the copious text?
Universal Truth #6: It's not the ink, it's the think.
The slide is not effective because there is less ink, it is effective because of the think that went into simplifying it and into extracting the story you want to tell. An obsession with having less ink is simply a forcing mechanism to cause more think .
Ok, time for you to practice a bit. Don't scroll further, just enough to take a look at this example… You've seen it at a major conference…
Consider our simple guideline: Declutter.
Get a post-it, sketch your version of it. As you consider the ink, let it force the think as well.
Ready to look at some alternatives?
You might have noticed the redundancy, the circles are trying to say the same thing as the numbers. Don't be redundant, unless it is your signature style thing. :)
If you just remove the numbers, you have something simpler…
It is not unusual for me to see tons of redundancy in our data analysis outputs. Focus on eliminating them. Picture a smiling Marie Kondo.
Universal Truth #7: Never force your audience to do the math. That's your job.
This is where the cute circles fail. Your audience has to figure out how much the second and third circles represent exactly.
Get rid of anything that fails universal truth 7. This would be my first step in decluttering…
Simpler. More direct.
Reduced clutter also encourages the application of universal truth 5, it forces you to see what's there. I mean really see.
When I look at the above, I realize the Analyst has ordered the story randomly. In the real-world Halloween comes first, then Cyber Monday and then Post-Holiday.
Small thing, but it matters so much to the flow of the story. This will ensure that in a very subtle way you don't jumble the brain of your audience…
Better. But, you are not done yet.
There is still too much text.
Universal Truth #8: Anything that is going to come out of your mouth should not be on a slide. (Also see truth #1.)
Applying the combination of what we did in the Ashley example, try and find a structure to communicate the story, and universal truth #8, you end up with this…
Started, Finished, Picked. Simple elemental structure to plant in the audience's mind.
Text simplified down to just the important days.
A little orange line to evoke the end of the period when retailers typically stop paying attention, stop doing the cool things and stop paying attention. But, they should not. The orange line subtly forces a closer look at the 44%, and that will prompt them to discuss what they can do to take advantage of this opportunity.
Nice, right? Scroll up a little, look at the original.
I'm sure you have even more ideas of how to make our two examples less cluttered. I would love to see them, hear your ideas. What matters most is less that my after versions are the gospelic best ones, I simply want to illustrate how with just a tiny amount of effort, and your brilliant think, the outcome can be so much better. You just have to care enough.
2. Don't fragment data, don't forget higher order bits.
It is not uncommon that we might end up presenting ten or twelve slides (or 106 if that is what you are into!). One sad outcome of this is that we'll tend to fragment the data, which results in our inability to tell the story as powerfully as it deserves to be told. It also forces a lot less think than might be optimal.
Let's look at this real data for a fictitious company… This is slide one…
Ok. Interesting trend. Up and down and up.
You ruminate on it, you let all the bars sink in. You are going to some kind of conclusion, and then comes the next slide with some more data…
It all looks new. You try to analyze it mentally (ignoring the speaker of course, yes they are still talking).
Usually the first instinct is to try and see if there's connective tissue between the two slides. It might take a second because of scale is different on both slides, but you'll ultimately get that the green bars are the same ones as before. They just seem to be different.
Why have the fragmented view?
Every time the ink creates a pause for you, an opportunity will kick in for more brilliant think.
What is the problem we are trying to solve with this data?
The higher order bit is that we want Reliance to be encouraged into action because as the above slides shows they are not competitive with the Market Leader. Sucks.
That in turn provokes these thoughts: Do we really need the trend if that's the higher order bit? Can we frame the problem in a different way, than percents, to create a sense of urgency? What above is really going to seem painful to Reliance?
Good think questions, right?
We are not really framing things cleanly, we are obsessing about something the Analyst loves, trends, and in the end, across two slides, the pain you want to deliver to drive action is completely absent.
Let's try something different, I would animate this to show one element at a time so that I could tell this story: In the last year there were 336 million (!) opportunities to engage with expressed intent, the average for your peer was 24 million…
Nice appetizer. Actual numbers. Making things really real.
Universal Truth #9: Actual numbers make reality real a million times better than percentages.
That set's the stage for one more click: Your performance doing this time period was…
Jaws dropped. People squirm in their seats. Things are getting hot. It does not look good. Sweet, that's what you want.
Everything you see above is framed simply, it is pretty much the same data you saw in the green/gray bars above.
Good so far, but the pain is missing. Let's show the cost by quantifying the impact due to the difference between 7 million and 24 million.
Be a true change agent, add the pain…
Now, you're cooking.
The other three numbers suddenly are a lot less important. the only one the audience can see is $80 mil lost revenue.
And, the conversation shifts from discussing the twelve month trends and the green/gray bars and excuses about advantages of the Market Leader. You have now singularly focused the conversation on the higher order bit: What should we do?
That's the difference between someone who reports data to someone who drives change in behavior. #winning
Yes, the trended data will be important in some executional kind of way. We have it handy, we'll use it. But, to make the most important decision, at a time when as a company Reliance is not doing well, that trend is simply a distraction.
Just to bring this home for you, here's my final version slide with the originals…
It is not just that we reduced the number of slides. It is that the fragmentation of data resulted in more ink less think and zero obsession with what really matters.
I do want to be cautious. I'm not fond of people who say that you should only have 10 slides, or some low number. It does not incentivize right behavior. I'm a firm believer in…
Universal Truth #10: Limit the number of stories you tell, don't limit the number of slides.
But as you go about that process, remember that fragmentation is not your best friend. And, remember to obsess about the higher order bits (what the heck is your end goal with all this data?).
Let's make all this real by looking at one more example.
This one pained me because the data is trying to make the case for solving a very important problem that women are faced with, yet the way data is being presented gets in the way of making a highly important point quickly.
Here's the first fragment…
It seems interesting on the surface.
Does not quite make the case for why it is an interesting something to ponder.
Here's the second fragment…
Also interesting in a superficial oh that is so intriguing kind of way.
Yet, even if here you can look at them both together, even with the y-axis being different for the two fragmented views, it is hard get the jolt you need to sit up and take notice. A part of it is also breaking universal truth #7.
Stop scrolling.
Go back up a little.
What would you do differently to make a version that is defragmented and keeps the higher-order bit squarely in the forefront?
Get a post-it. Sketch something.
Ready?
[update]
Rishi Shrivastava shared this waffle chart as another lovely way to illustrate the data simply…
It is a different way to bring focus to the conversation as is clearly visible from it's strengths. For my take on the original, continue reading…
[/update]
After thinking about what the higher order bit is, to help me focus sharply, and applying the other lessons you've learned thus far, here's my simple version…
Sharp focus.
The biggest delta between the two trends is for women under 40. That's the only thing I care about.
The moment you see those two bars, so starkly illustrate the problem, you are on to what the heck should we do about this?
I felt this still did not "starkify" the opportunity clearly enough.
So I went and got me some real numbers…
Boom! Excitement delivered.
We often break our data story into too many pieces. Partly because we have this hidden worry that our jobs are not that justified. We throw off a lot of data as a subtle way of earning a great job performance review. Sadly the outcome is exactly the opposite.
What you want to do instead is to do all the slicing, dicing, segmentation, beautiful math, and then step above it. Reach for the higher order bit. It will lead to defragmentation of what you are trying to say, and the ensuing simplicity will drive the type of impact that will earn you a promotion faster.
Universal Truth #11: Data is just a means to an end. It can't be the end in of itself.
3. Obsess with deleting information provided.
This rule originates in the above universal truth. Because we are trying to get to the end that will drive the higher order bit outcome faster, one strategy you can embrace with impunity is to simple delete stuff.
Train yourself to find joy in killing things. Clipart of course, you are not a child after all. But all kinds of other stuff too. Enjoy the killing.
Whenever you see a data output, your first instinct should be what can I kill in service of enhancing simplicity and the effectiveness of the story?
Take this example… Don't drink it all in at one time, slow down, let it all sink in…
What do you see?
Grab that post-it.
How many universal truths does the above slide break? How many rules we've looked at so far are not followed? How many things you personally have in your arsenal of things you hate that are above?
Ready with your rough sketch of what you would do?
My approach was simple. Delete anything that's redundant, and simply visualize what's left for sharper focus. Here's my version…
Honestly. What else is there to say?
Show the distribution, add the ages, use colors to highlight segments you want to bring attention to, add a orange bar so they don't have to do the math, and happy birthday!
You get to this important place so much faster: What the heck am I actually trying to say?
Scroll back up and compare.
Let's make this even more real by looking at a more complex example of a multi-channel marketing recommendation…
I am sure your instinct is: OMG!
There is a lot going on here. Everything seems sub-optimal. But, what I want you to recognize and appreciate is that there is no way that the person who put this together did not do a ton of work to get to this point. But, it dies at the last mile.
If you assess it carefully, the above recommendation set is trying to do five different jobs. That contributes to the problem. In the end, the audience is not impressed by the intelligence, they either start swiping left or right on their phone or become super confused.
Grab your post-it.
Just based on the information you see above, what would a version you'll simplify look like?
Apply the lessons learned thus far, primary amongst which would be to just delete, delete, delete.
If it helps, ask yourself if I had to focus on this data doing just one job, what would that one job be?
Don't scroll. Do the work, I promise you'll learn more.
Ready?
There are many valid stories to tell above. My decision was that the most impactful one was the recommendation around the media plan for the six month period. Additionally, the higher order bit to solve for was to give that media plan a purpose.
Here's the quick five minute simple result…
Nothing gloriously sexy. Just effective.
Shows the key time-periods, start and end dates of each media recommendation, and leverages the See-Think-Do-Care business framework to give each media effort a purpose. The latter is especially important because it directly ties to what content the ads/marketing should contain, what the tone and texture should be of the landing page/app experience, and what we'll use to measure success (S, T, D, C metrics).
Universal Truth #12: Never ever never never end your story with a buffet of purpose-deficient "ideas."
Notice the subtle shift in posture as well.
I dislike ending on a buffet of choices after all the hard work. Here are five ideas, take whatever you want. It is is irresponsible behavior. You did all the work, why not benefit the audience? Hence, you see everything framed into two clusters of intent.
This does not mean everyone will agree with you and lift up your chair and parade you around the tiny cramped conference room. They'll argue with you, they'll debate amongst each other, there might even be a productive food-fight. At least everyone will know what your POV is, and over time your profile will be bigger and more influential.
Ok. Exercise time.
Here's a simple example similar to the before versions we've looked at in this section….
What would a version fixed by you look like?
I bet it would only take a few minutes for you to apply the lessons in this post. I would love to see what you do with it. You are welcome to comment below, or you can email me your version.
In a week, I'll update this post with my simple five minute version and I'll share a couple examples from you (with your name, and permission).
[update]
So many exciting ideas from our readers, I was truly inspired.
Jordan Zivoder's version used a very different approach to visualizing the data…
Each strand is easy to follow, you can more easily see the outliers.
Roman Mozil's version converted the original into one that uses a very light touch to bring simplicity…
Even though the amount of data in Roman's slide is almost as much as the original, it does not look like it. Right? A testament to the power of spending just a couple minutes thinking.
Paul Berney's take on the slide was to pick a key point, and then illustrate it with an image…
In my work you've seen I stay away from stock photos. In my experience they are massively abused, and if I have 10 mins to spend on presenting data, I want to spend all 10 mins on my data and answering so what. But, in this case you'll note how Paul fades the image into the background and leaves the focus on what matters.
Jakob Samuelsson and Hugo Galvenius put their thinking caps together and came up with this really simple version…
Hugo and Jakob wonderfully demonstrate the value of picking just one story to tell.
David Solito's version is very stark, focused, and expands the think…
He made some assumptions, and tied the difference in behavior to outcomes. His second slide…
As you can imagine, this should drive the key points home very effectively and drive action.
Here's my version of the original that I'd shared in the storytelling ideation session I lead (where all these examples came from )…
My goal was to narrow down to just the segment that has something interesting to talk about, 18-34, and demonstrate their behavior starkly. Jina Chan also had a version close to the one above!
This slide was followed by a slide where I answered the so what question by sharing ideas of what USCO should do about this trend.
All the examples above prove there are many ways to tell the story simply and I hope you are inspired by the efforts of our wonderful readers.
[/update]
Oh, and one more universal truth from the macro thought in this section…
Universal Truth #13: An Analyst is not a dispassionate participant. Invest in knowing the business, develop you personal point of view. Then, Express It. Let the audience benefit from your wisdom.
4. Don't run away, make the tough choices.
I'm sure you are noticing how these rules act like layers on a cake. One on top of another on top of the next one… each adding yummy deliciousness.
[In case you jumped here, just a reminder that was at the start of this post: All the examples are real, from different places, but names of companies and marketing channels have been changed or anonymized.]
Here's our next example… don't scroll too far ahead, grab your post-it again, note down what would you do differently…
Some things are obvious. Yes, redundancy can be eliminated. Yes, colors are a bit garish. Yes, text can be optimized. So on and so forth.
I hope your review also exposed that there are two different ways to frame the story. The author above decides not to make a tough choice. I encourage you to make tough choices. It is hard enough to change minds in the first place, harder still when you are sharing lots of data, why make it nearly impossible by being shy.
Since this slide is framing the value of AOL, we could frame the value as an absolute or with the percentages. The problem with raw numbers is that you don't have any context to make them a bit more hard hitting. Hence, my choice was to anchor the story in Reach. You'll see in a moment that it is an imperfect choice, but it is important to make a choice.
Here's the first view you would see in my version…
What I want to communicate is my assessment of what is AOL's core value proposition here, something that will surprise people.
The second click would revel how the story is framed… For folks that unfortunately suffer from Diabetes, AOL has close to 50% Reach…
I could just keep it at the Reach number, but you've noticed an aversion to just percents in this post as they don't quite get the message all the way across. The number in the yellow font adds a little bit of sizing of the audience, invaluable context.
A couple more clicks, and the rest of the story fills itself out…
You can get the core message very simply now. And, you have a sense for the size that is easy to understand.
The challenge of course is that it is quite odd to see that the 2,120k number is on a graph that is visually lower than 106k. It is hard for folks to wrap their heads around that concept – understandably so.
There is a legitimate case to be made that you could cross-tab this simpler version to create a workable alternative. Have the orange line be a raw number benchmark, say 500k, the blue bars represent the raw numbers, and yellow text becomes the Reach numbers.
But, it was my judgment, there's that word :), that based on my understanding of the situation that a Reach centered story would have a higher impact in terms of the idea that we were selling.
I could have totally bombed, it is a price I'm willing to pay. I'm doing it to have the ability to simplify things, and take risks. I'm willing to make a tough choice. I highly recommend that you do too. Your couple failures will be dwarfed by the numerous successes.
Let's try to apply this rule one more time.
Here's the picture that was painted… It was a reflection of a company's possible growth…
What's missing above is the Nascar-esq parade of logos, to preserve anonymity. You can be assured that it was appropriately eye-striking.
Get a post-it.
If you had to make tough choices, what would they be?
Sketch it out.
Some of the rules you've learned thus far are easily to apply, and I hope you caught those things.
For me there are two tough things to consider. Should we have the logo parade, though it contributes in the above view to explaining each bar? Is this the best way to frame the story we want to tell?
The first one was an easy decision due to this…
Universal Truth #14: Don't tell the audience things they already know.
Please remember this. Every moment you have is precious. If you don't have enough to say, it is ok to fill the gaps with thoughtful pauses. Regurgitating a parade of data/ideas/facts that people already know is a terrible choice.
On the second question I decided to step away from the waterfall chart. First, because I'm not sure it is adding much value. Second, you may or may not have noticed this, the second bar, 1.9 M, is actually a loss. It is illustrated weirdly, I wanted it to be more clear.
Here's the first view I would show… The past and the part of next year that is not great clearly identified…
I love having the focus starkly on the core of what we want to say, and all the negative space to let the data/visual breathe.
Then, element by element we reveal rest of the story…
The green and the yellow bars have insights that the presenter is delivering, we put them clearly out there for something new to be told to the audience.
Compare the before and after, I hope you feel the shift from having the slide/data being the center of attention in the original to the person telling the story becoming the center of attention.
One more exercise to help you truly internalize the lessons.
Here's the original that you can take a crack at fixing…
What would your version with tough choices look like?
Make tough choices. I would love for you to email me your version. Or, share your thoughts in the comment section below.
In a week, I'll take a couple of the best versions from you and add them here with your name and permission. I'll also add the version I've already created
[update]
This was a challenging assignment. Bonnie captured it perfectly in the note to me: "Your funnel chart concerning purchases is way too confusing to even guess what the main objective is or what numbers go with what." :) Absolutely true.
It shows how the very best Analysts, at companies with great data environments, wonderful tools at their disposal, create things whose only outcome is a brutal death at the last mile.
Here are some brave souls who still gave it a go!
Andrew Clark took up the immense challenge and created this simpler version…
This version allows us to understand what was on the original a lot better. (What Andrew had zero way of knowing was that all the data on the slide was for Footage, none for Image purchases. The original is just so confusing!)
Hugo Galvenius and Jakob Samuelsson took a crack at making sense of the confusion in the original…
I love their taking everything down to one simple point. :) Really good.
Roman Mozil's version is so, so, much better than the original. He's picked one story to tell, and then told that story really well…
I appreciate his addition of a little text with a small explanation (you'll notice I dispense with it, but Roman did it better).
Here's the version I'd shared at the storytelling ideation day…
I picked one story, like Roman, and then used the visual treatments that made the story as simple and as attractive as I could make it in five minutes. I did not even do a graph, you'll notice it is just a table in Excel. The image was take with the table edges, they serve as a rough scale. Just to communicate that, I have only one number, 38%.
[/update]
Remember, life is too tough for you not to be brave and make tough choices. :)
5. So what? So What?? So WHAT!
We have one more rule to go, but I think this is probably the most important one on the list.
One big way to focus the story you are telling is to ensure every beat in the story has a purpose. In our case, every table, every slide that comes from a piece of data, has to pass the so what test. It should be really clear why someone should deeply care.
An example of this that you saw above was quantifying the pain. The so what was $80 million in missed revenue. It can be other things as well, you saw the quantified opportunity size above. It can be any of the numerous brand metrics avilable to us. But, it always has to be something that would give your table/data/slide a clear purpose, something that causes people to lean-in.
Here's an example that illustrates that….
If you've been paying attention you'll notice this is a great example of how you can be seemingly uncluttered, and yet miss the mark by a wide margin.
And, if you've been paying attention you'll quickly get rid of the disappointing clipart, you'll notice the redundant numbers and text, you'll pick out the sadness of using percentages of percentages (see universal truth #7 above), additionally you'll be confused about what the +44% YOY numbers refers to because nothing seems to match that number, and other sub-optimal things.
You'll also consider the higher order bit and realize you can make tougher choices and frame what you want to say simply…
Sweet, right?
Gets the very core of the message in front of the audience, as fast and as simply as we possibly can.
But, you are like… Avinash you said quantifying the opportunity size is more powerful than percentage deltas!
Yes, you are getting ahead of me.
Here… Yes, that is always better… quantify…
So 22 million more Hispanics watch videos as a part of their shopping experience.
You see a compelling so what from the data there. Video is a valuable source, more people are using it, hurray.
I want you to push yourself harder. Shoot for the very best so what. In my experience, that comes from obsessing about the impact.
In this case…. Here's a more powerful so what that we found with a little bit of effort… People watch videos during their shopping experience, fine… What's the influence of that video watching?
OMG. Great question….
This last piece of data, how much video advertising's influence has changed between 2013 and 2015 in the individual making their final choice, is killer. You're not just talking about raw reach, raw views, raw acquisition metrics. You're talking about tying the acquisition opportunity to the outcome/impact. That my dear readers drives action. The ultimate destiny of data.
Universal Truth #15: Purposeful data wins. Does your data have a powerful answer to the question so what ?
This is the hardest possible thing to do. It requires a fundamental understanding of the business, the various business strategies currently being executed, merging of various data sources, deep analytical skills, and an ability to predict impact. It is certainly the place where Analysis Ninjas are separated from Reporting Squirrels.
6. Shift your worldview to your client's.
Regardless of your role when you email a succinct summary of your insightful findings, or when you are standing in front of an audience, you are selling. Your audience – internal, external, at a conference – are your clients. Bring that perspective with you. It will cause you to lean-in a bit more.
Here's an example of some of the rules you've already learned, related to simplicity, tough choices, with an additional posture I want you to adopt: Taking the client's perspective.
We look at things too narrowly, or in a sales situation (see above) only care about ourselves. It is natural. Look at the data, think about the story, step outside your body, look at the situation from your client's perspective.
Let's make this real.
Ponder this incredibly complicated example…
What are the issues you see?
Take a post-it. Make a note of them.
There is a buffet here, we covered in story three why buffets are almost always sub-optimal as the client/audience does not benefit from your wisdom. You'll also see that there is just too much stuff, some of it unnecessary and other bits expressed in a more complex manner than might be optimal. It looks like there are two choices, but they are hard to see.
Applying the lessons we've learned today, here's what you can get to with only a tiny amount of effort.
If you are going to have a title (honestly, they add little to negative value), try and put some attitude in them. :)
We center the story on two choices, and make it super-duper easy to compare every equivalent element across the choices.
It takes you maybe five seconds to see the deltas, and perhaps make up your mind as to what you want to do.
You could be happy at this point. You should not be.
If you go back and look at the original you'll notice that that you don't really have two packages to choose from. Scroll back up, it is not very hard to spot.
Two of the elements are the same. There is a new element in package two, it is of immense value to consider what value you'll get from adding it. And, is it really worth giving up 600k views for $56k when it comes to TrueView In-Stream Skippable?
That should lead you to wonder, as did I, is it really of value to get "3.0 M impressions" for $100k with Pre-Roll, when compared to "2.8 M Views" for almost half the cost with Bumpers?
If you truly think about what you are saying from the unselfish perspective of your client/audience, because you loooooove them, truly and deeply, you might end up framing the opportunity completely differently.
My choice was to eliminate the poorly created package business complexity, frame the story as three advertising choices with the investment required for max impact, and then, with truly the client-centric view, say that they should choose just two…
Because I don't like buffets and want the audience to benefit from my point of view, I put two smiley faces for the choices that I would make if I were them.
Yes, I can get more money by choosing the first two. But, I would value Views more than Impressions. Hence my final version above.
I leave some money on the table in the short run, but I leave the lasting impression that I care for them first, I care for myself fifth. The true and glorious path to both driving big change and creating trusting relationships.
Always, always, always, look frame things from the client's perspective. [If you are in Team SDM remember we have even higher standards: Obsess about the client's client!]
This is a topic with deep fascination for me, if that's true for you as well then here's a small collection of paths you can walk down in order to truly master this delicious topic.
There are fifteen additional rules, with tons of real-world before – after examples like above, across these two posts: Great Storytelling With Data: Visualize Simply And Focus Obsessively | 7 Data Presentation Tips: Think, Focus, Simplify, Calibrate, Visualize
If you are interested in inspiration related to different types of visualization techniques you can use in your analytical practice, you are going to die of delight: Data Visualization Inspiration: Analysis To Insights To Action, Faster!
The very best lesson when it comes to data, the one thing you should never forget and you can never apply enough of, is this one: A Great Analyst's Best Friends: Skepticism & Wisdom
At Google we've developed an amazing framework we use to tell stories with data, more on that in this Forbes article: The Digital Evangelist Leading Google's Storytelling Movement.
I hope you find the knowledge in these links to be of value as you tell your stories with data.
As always, it is your turn now.
Have you shared your versions of the two exercises with me? Which of our six rules is your favourite? Are there other rules you would apply to any of these examples to make the output even more impactful? Which after version did you love the most? Did any of your sketches take you in a different direction than where I ended up? Do you have golden rules for telling stories with data that are not covered here?
Please share your insights, critique, lessons, examples, post-it pictures, and guidance via email (blog at kaushik dot net) or via comments below.
Thank you.
For Reliance's Yandex Search, I believe, they wanted to show where they stand in 2016 (currently) against 2015 (lets say, when they started).
Hence, they might also like to showcase that click were initially less than 2M and now it is 7M. (also, increase in revenue)
Arpit: I agree that it is an element that might be of value, and in choosing it is very helpful to have the context of the meeting. You'll pick the element that most clearly solves for the higher-order-bit. In this case that's the reason the 2M dropped off as not necessary.
In every case though, you'll decide to choose the minimum most effective data presented simply.
Avinash.
Hi Avinash,
They are all excellent thank you for sharing your insights.
For me, number 1 (including the universal truths!) is my favorite.
I will never forget creating a dynamite PPC attribution report for a healthcare CEO with a 6 figure monthly spend and because the font was too small he refused to read it. Talk about deflating.
You said:
"That's the difference between someone who reports data to someone who drives change in behavior. #winning"
This really resonates with me. Data that doesn't help people make educated decisions or drive change is just 'number on a page'.
Respectfully,
Alan
Alan: You are incredibly sweet to share your story, I know it will resonate with our readers (and also inspire them).
Your latter thought is the reason I've framed to the teams I've had the privilege of leading that there should be three components to their output. Insights (data) – Action (specific to-dos) – Business Impact (predicted outcome). Most analysts only do the first.
Avinash.
: )
Just in case others missed that:
Data > Action > Impact.
Thank you, Avinash.
Great stuff here.
I'm a big fan of trying to simplify data in a way that is actionable/meaningful to team members/departments outside of the analytics team, but sometimes I struggle to find a good balance.
You're absolutely right that a ton of work & resources can go into collecting, organizing & displaying data ..but the end result doesn't mean anything to anyone.
Kiyo: My strategy to strike the balance is rule #6: WWCW
What would the client want. :) What's in the client's best interest. It helps cut a lot of the "crud" out because I'm applying the Occam's Razor to delivering their desired outcome.
Avinash.
There were so many times in reading this post that I stood up and screamed YES. I'll have to return a couple of more time to understand all the layers you applied. It seems so easy on the surface.
The example I liked the most was the last one. Your reframing of what was being said was clever.
I'll share my ideas via email. Thanks for this fantastic post.
Avinash, very nice article and cool examples. It is definitely going to be one of my favourites.
I always follow the same philosophy you advocate, although there are handicaps I face sometimes. How can I fight against the concerns that comes from an audience in the meeting room over simplicity of the analysis/story?
Very often heads are craving for details and those meaningless trends. I also sometimes think that people, including high-rank managers, focus on the wrong things.
Do you have any advice on how to address demands of those involved into discussion of one or another topic?
Igor: It is hard to answer this with a pithy answer, but let me shine a light on one higher order bit.
You have to figure out what business you are in as an Analytics team, and then through your execution reset expectations of your management.
So. If the business you are in is to provide data…. It will influence how you think, it will influence how you execute, it will influence the skills your team has, and it will drive the outputs you provide. It will be almost impossible to get your Management to expect anything from you other than "lots of data." Even if sometimes you try really hard to show you can give A and BI part of the I-A-BI model.
If business you are in is to provide ideas to change the business, your focus will be different, the skills you'll hire will be a little different, your execution strategies will be different and the outputs you provide will not be I, that'll be there but a lower order layer, rather it will primarily be A tied to BI concretely. Your management team will expect you to only come to them with ideas to action. They might want to see some supporting data, but over time they will develop confidence. The business impact you share for each action will help prioritize discussion. The meetings might still be 60 mins long, but the time will be use arguing about actions and assigning ownership for each action – rather than arguing about the data.
Here's how you know you are in the second business: The governing document that drives your existence will be the Digital Marketing & Measurement Model. It cannot exist if your Management thinks you are in the data business. Or if not the DMMM, your team's existence (reports, analysis) will be dominated by a constant obsession about Outcomes.
This is not easy to do. But, it is doable.
I hope this is of value.
Avinash.
Me too! My managers always want complicated charts with trend lines and loads of historical data (because that's what THEIR managers want to see).
How should you respond when the audience WANTS to see data-pukes about low-value metrics?
(Wonderful article. I'll be revisiting it several times in the coming months. Thank you so much!)
Jessica: The only way out is to figure out how to "move up the supply chain." My broader point to Igor above.
Maybe you can do this, I've shared it in the past here, give them Bounce Rate and Page Views… but also include one recommendation, just one, for action they can take based on a better metric Page Value. That get's the ball rolling in the direction where both you and the data are seen differently. You move up the supply chain. :)
Avinash.
This article is very helpful for Analyst like myself who is pressured to keep churning data out. I have a collection of ideas from you that I can use to push back.
I'm sure you hear this enough but please keep sharing your valuable knowledge with us. There are few sources in our space that provide this level of intense real-world learning.
Thanks Avinash.
Heyy Avinash!
Such simple but powerful suggestions to clearly send the message across instead of a bunch of data points.
Loved this post. Thanks for sharing.
So. Awesome.
As always Avinash. Thank you.
Thanks for a great article!
Loads of takeaways beyond just the data visualization that I can apply in how I think about my role. It is as much of an attitude shift as an analysis shift.
Awesome article.
I've picked up numerous new filters I can apply in my monthly executive report-outs. Currently I am guilt of many of the sins you have outlined in the before versions. Loved your universal truths!
Thanks for share bro.
I love this problem, thinking, solution format. I have a collection of Post-Its next to me with scribblings of my ideas as I went through. It was useful to contrast them with your after versions (you are significantly more ruthless in your versions!).
I have four other tabs open with your bonus reads. I think you might have already written enough to create a book on this topic Avinash. :)
This is amazing!
Thanks again for giving me so much to learn.
I echo the sentiment others have expressed, this is an invaluable collection of knowledge.
Articles on this topics is either inaccessible with the amazing infographics that people tweet from Info | Beautiful or theory that is hard to internalize. The real-world examples you've sourced to help make the key points really bring the lessons home for me. It is content that it accessible and will have an impact.
Many thanks for sharing!
Hey Avinash,
Thanks a ton for this post. I am gonna print it and pin it on my desk to refer it again and again.
However i could not resist to put a case in place, when audience is shown this kind of insights (which is suggested in post) and questions like how did you reach to the conclusion, from where have you got this come up and we end up discussing CDPs or Data pukes. Then whole discussion goes around on proving the sanctity of the analysis/insights.
So isn't it right in putting a disclaimer to above universal truths that those will work only in a trusting atmosphere.
I would love to you here your take on this.
Thanks.
Mudit: I appreciate you making the important point.
Usually, if we are invited in the room with our Sr. Management there has to be a level of trust and expectations that we will provide recommendations to drive change. We might be fine in that case, if we deliver the goods.
For our other initiatives, where we are meeting needs of various org levels it would be of immense value to follow your cautious advice.
Avinash.
Hi Avinash,
nice post, and so truth….
Avinash, you've done it again. If there were nobel prizes for web analytics you'd be due a third or fourth at least….
Thank you for this. Thank you for all the effort that went into explaining it so well.
Have you considered teaching courses on this where folks can practice and get feedback?
Melinda Byerley
Founding Partner
Timeshare CMO
timesharecmo.com
Hi Avinash,
It's Javier, from Spain. I'm a brand new data and analytics interested. I just bought your book Web Analytics 2.0. and read this whole article… hope to learn much more from you along the following months!
Just a question, I'm 22 and I currently work on SEO. I'm really interested in working with data (I'm learning a lot about Google Analytics), do you think that it's possible to work in SEO and Analytics at a time or it might be too hard?
Thanks in advance!
Javier
Javier: I encourage you not to get too narrowly focussed in your worldview just yet. You are young, you have lots of energy (and time! :)), it would be optimal to consider focussing on two or three inbound marketing strategies rather than just narrowing to one.
For example, you could challenge yourself to be fabulous at Search (SEO + SEM). You've just made your life exciting, and yourself very marketable. Or maybe SEO and Email. You catch my drift.
You are never going to escape being good at Analytics, literally nothing is possible on the web without you using data. You don't need to specialize in Analytics – as in do the types of advanced analytics that is often in posts on this blog. But, you'll need to be good at all the basics, picking good metrics, custom reporting and a little bit of segmentation.
I do not believe that would be too hard. In fact, it will cause you to stand apart from the competition.
Avinash.
Avinash, I am a huge (Super huge), fan of yours. (After reading Web Analytics 2.0). I read your recommendation on SEO. I wanted to know if you recommend anyone other than Jennifer Grappone's book. (Because I brought that). I was reading through one of your projects, Market Motive, and saw that Danny Dover is SEO teacher on that. I wanted to know your views on it (I apologize if I am putting you on some spot here). I take your recommendations so seriously that I have purchased almost every book that you have recommended so far. (And logged onto my twitter after a month, just to tweet you)
Any insight that you give me on this, will be very helpful. I am pretty unhappy with the traffic I get from Search, so if you could suggest me anything else, I would be happy.
I have learned this from your books — it is important to segment things. So, I am segmenting between books, tutorials, experiments, et cetra.
Thanks in advance for the answer.
Deepak: I appreciate the kind words, thank you.
Danny is an excellent SEO, and his course is very good (and just $99 for 180 days at the moment):
https://www.simplilearn.com/digital-marketing/seo-basics-training-course
If you were only looking for a book – that solves a very different problem than Danny – You can consider SEO for Growth by John Jantsch.
Re your low traffic… Consider multiple strategies (say SEO and Email), and of course ensure that your content meets this criteria: Incredible, of value.
All the best.
Avinash.
I wanted to commend you on the stories you wrapped into each recommendation. I was so excited to see each after version well before it came up. I Googled OAB and got wrapped up into each example.
This post has great stories about how to tell stores!
Thanks avinash.
Thanks Avinash for posting this great article!
This is the difference between knowing how to report-make conclusions out of data and having the gifts to transform data into stories. And this skill is powerful in communication-teaching and of course in analytics. You should consider launching a course about this.
A liked the "declutter" mantra.
Sorry for the newb question, but what do you use to create your data visualizations? Excel just doesn't get the job done for me.
Brad: This post has some examples of tools I end up using:
Data Visualization Inspiration: Analysis To Insights To Action, Faster!
https://www.kaushik.net/avinash/data-visualization-inspiration-analysis-insights-action-faster
I have to admit that for business meeting communications, having something like Excel or Tableau bind my hands, because they can only do simple things, is a pretty good thing. It forces me to think about what's the most essential thing I need to say. And, then show it simply.
Avinash.
Portraying your data in an appealing way will provide you with a valuable, visual piece of content, that you can share in email updates, reports, and other content.
Great post!
Another fantastic post; the more I read your blog post, the more I realise how little I know and how we all have a duty to move away from blah blah and focus on actionable insights.
Thank you for sharing your expertise and getting us motivated to think deeper and support others in making better decisions.
Hi Avinash,
It was a great idea to involve us. I liked it. Great challenge!
I hope this experience will motivate you to reiterate the experience.
We can learn so much from others.
Thank you.
Hi Avinash,
Great article. When I read this post I get the feeling to differentiate to use data visualization in two ways.
The first for my own or people who are "in the topic". For these people are more informations necessary to get a better overview. If you cut to many informations you don't see all the important "side effects".
The second one for people which are not very deep into your work. For these people you should keep everything short and simple.
I think it makes sense to create different data visualizations for different target groups.
Thank you,
Tim
Tim: You are right, it is important to differentiate between audience needs. Perhaps, it is will help to re-frame this using purpose.
This post, and guidance, is for people who are in contexts where the purpose is to communicate a key insights with specific recommended actions. You are trying to get people to do something based on your intelligence applied to the data.
You might also be in contexts where the purpose is to share data. Someone else is responsible for drawing out the insights, then identifying what the recommend actions might be for the business. Your job in this case it to use your intelligence to find all the data that might be needed, stitch it together, and share the data.
Once you define purpose, you can use your time appropriately (and also decide if anything needs visualization or not!).
Thanks.
Avinash.
Great collection of ideas.
What makes your blog distinct is that you are sharing tips that are based on what people really do in their job. I am able to apply most of these in my day to day job.
Thank you.
I'm just now coming across this post and wanted to thank you for sharing.
I will be referencing this post whenever I'm putting together a story based on data.
Thank you again as this post has motivated me to think differently.
Very good article with very good ideas and perspectives.
The one we have to consider more deeply in my agency is #5. The structure of our engagements means that we have little time or even incentive. On the long run your stress on this is important for us to figure out.
Thank you.
A great collection of ideas, I learned lots of new think with this post Avinash!
We know that a lot of data is not great yet we seem to add more and more stuff until there is no white space left. I'm going to be vigilant about this in my work.
Thanks for constantly setting new standards for us.
Very interesting article with a lot of useful information Avinash. Thank you.
The call to arms against "crapification" needs to be made constantly. There is a lot out there.
It was interesting for me to read through the implications of not fragmenting data. It is hard to do as we consistently want to drive deeper. Good reminder.
Hi Avinash,
Thanks a lot for this amazing article ! So much to learn from your experience. I was also wondering about the tools you are using on a daily basis to make that kind of visualization. Do you use a tool in particular ?
Thanks a lot,
Thomas
Thomas: Some of the tools I end up using are in this post:
https://www.kaushik.net/avinash/data-visualization-inspiration-analysis-insights-action-faster/
Other than that, day to day I find the limitations of Excel wonderful as they force simplicity. Beyond that, I'm a big fan of https://d3js.org
Finally, more tips in this blog post (along with some tools in the comments):
https://www.kaushik.net/avinash/data-last-mile-gap-visualizing-for-impact/
Avinash.
Hi Avinash,
Recently came across this post, and was indeed a fantastic learning curve.
I am trying to mimic the ideas presented above to convert a 8 page report document with lots of tables and few graphs and text to a single DASHBOARD whilst still creating links to other report/tabs for a detailed version of that 1 particular "Tile" used in dashboard.
Struggling to do so, but hopefully will get somewhere.
Heyy Avinash!
Such simple but powerful suggestions to clearly send the message across instead of a bunch of effective points.
Thanks for sharing.