The Back of the Napkin Part 11

To see why we should make the spending decision one way or the other, we're going to have to look at our entire industry: who our compet.i.tors are and their growth projections, how customers and sales trends are changing, and how changes in platform technology will impact revenues. It's only through seeing all that information tied together that the picture we need will emerge. But how can we see all that? Is it even possible to plot together that much information in a meaningful way?

The codex tells us that the coordinate system of a multiple-variable plot is, by definition, composed of three or more variables. Here we have five or six potentially meaningful variables, so let's go ahead and see what happens when we superimpose them onto a single picture. We'll be drawing an elaborate, quant.i.tative, visionary, comparative, as-is, and could-be plot, a window into the closed box that is our industry. If we can open that window, it should give us a persuasive visual argument for why we need to spend the money now.

REVIEW: A MULTIPLE-VARIABLE PLOT SHOWS WHY.

After we'd seen who, what, how much, where, when, and how, we saw reason (or reasons) emerge. The longer we watched everything interact and focused our attention upon cause and effect, the more we began to understand why things worked the way they did. In order to show others the reasons and to begin to make predictions about how things will work again, we create multiple-variable plots.

Chapter 5 told us that we see why when our mind's eye combines the other ways of seeing. To create a multiple-variable plot we do exactly the same thing, only this time combining them all on a sheet of paper. We start with who/what, work through how much, shift to where, and then add in when. Because we've already sketched similar drawings in the previous sections, creating this plot will largely be a review, but with two big differences: First, we'll be layering everything into a single picture rather than separate pictures, and second, we won't start the who/what with a portrait of our customers, we'll begin instead with a portrait of our compet.i.tors.

Multiple-Variable Plots: General Rules of Thumb 1. Multiple-variable plots aren't hard to make, but they do require patience, practice, and, above all, a point. Of the six frameworks and hundreds of picture types out there, a well-thought-through and clearly drawn multiple-variable plot is the most powerful and insightful we can create. (We'll talk about why that's true below.) That said, I can't recall ever seeing a simple explanation in a business book of how to draw one. My advice is this: Begin with a simple x-y plot, using any two qualitative variables for which you have data as the two coordinates (remember, if they turn out to be useless, you can always change them later). Plot in any quant.i.tative variable for which you have data using appropriately sized bubbles in the middle, starting with just one point in time. Then add another set of bubbles showing the same quant.i.tative variable at another time. That's it-all you need to complete a multiple-variable plot either as a final picture or as a launching pad for adding more and more variables.

2. Medium-thick soup is best. What a multiple-variable plot really does is to create a scale model of an entire business universe or business problem. When we create one, what we're hoping to do is identify a limited number of aspects of our industry (or problem) that may have great influence on one another, so that we can pull out just those and look at them side by side without the distraction of all the other variables out there. Too few variables and we end up with a simple bar chart-useful on its own for many things, but not for developing real insight. Too many variables and we're back to the original problem of too much to look at and we haven't accomplished a thing. Again, the only way to know the "right" number is to start plotting and see when useful ideas emerge.

3. Anything can be mapped to anything else, but... The biggest danger of multiple-variable plots is that because they invite the layering of many data types, they can make it too easy to "discover" connections between variables that actually have nothing to do with one another. This is the great challenge of statistics and even basic science: keeping "correlation" (the appearance of similar trends between different variables) distinct from "causation" (the direct impact of one variable upon the other). While it may be tempting to map global temperature fluctuations to the frequency of Baywatch reruns-with very possibly a high correlation factor-it does not mean that one necessarily causes the other.

Back to SAX Inc. In our industry, we face two categories of compet.i.tors: the old guard (that's us-SAX Inc., along with SMSoft and Peridocs, companies that we've competed with for the past decade) and the new arrivals (Univerce and MoneyFree, which just appeared on the scene a couple years ago). The two groups are further differentiated according to other specific criteria: We big three have all been in business for at least ten years, have all built our software on proprietary code and platforms, all offer software with lots of features, and all make our money through the sales of our software products, throwing in the upgrades and service for free. The smaller two companies built their software using open-source code, have few features, and make their money from support contracts only: They give away their software for free, then charge their clients for upgrades and service.

Portrait of our compet.i.tive set, representing two main groups and differentiated by age and differing approaches to the market.

That's it: five companies, two different platforms, two different ways of doing business. Now let's look at a simple numeric comparison to see how much revenue each of these companies earned last year. As we map out the companies by size (using proportionately sized bubbles to represent revenue), another trait emerges: The old guard made all the money last year, while the new arrivals barely made a dent. SAX Inc. lead the pack with revenues of $25 million, followed by SMSoft at $20 million and Peridocs at $18 million. Univerce came in at $3 million and Money-Free made a small blip at $250,000.

Now let's look forward. Using a.n.a.lysts' reports, Wall Street projections, and the industry rumor mill, we can project what revenues are expected to look like among these same companies at the end of next year. We already know that our sales are flat, but here is some new information: SMSoft is in negotiations to buy Peridocs, which will create a combined company with projected revenues of $40 million. On top of that, a.n.a.lysts predict that Univerce, a company that didn't even exist three years ago, will surpa.s.s our projected $30 million by more than $1 million, knocking us from first place into third. Even puny MoneyFree will likely bring in $18 million. What?!

That's a lot of industry change in a short period. Aside from the big merger, what else could be happening? Obviously, there's more going on than this simple how much chart can show. We need to not only see how big these companies are, we need to see where they sit in relation to one another according to customers, platforms, technologies-all those unique variables we identified in our portrait. What we really need is an industry map.

Our compet.i.tors' revenues as projected at the end of next year.

We begin our plot with the horizontal coordinate, in this case type of software platform, then add the vertical software features axis.

Let's try it. Let's plot together what were otherwise separate pieces of information and see if connections do emerge. The specific pieces that we want to see together are things that we already know: compet.i.tor name, type of platform, range of software features, revenue, and time. Remember that a multiple-variable plot overlays three or more different criteria, and to get started we just have to draw in one or two initial axes and give them names. For example, proprietary standards versus open standards plotted against full features versus few features.

Now that we've got an initial coordinate system laid down, this picture becomes like any other landscape map, and all we have to do is draw in the features. Since we've already got the bubbles representing last year's revenues ready (our third variable), we can place them in the areas of the plot indicated by the coordinates. For example, SAX, SMSoft, and Peridocs all slide to the proprietary side while the others slide to the open side, and vertically all are arranged according to number of features (SAX has the most, followed by SMSoft, etc.).

With our coordinates mapped in, we then draw in the features: in this case, the spatial locations of ourself and our compet.i.tors.

So far we're not seeing anything that wasn't already captured by our mind's eye: The big bubbles (more revenue) have more features and are based on proprietary platforms, last year, anyway. We didn't need the picture to tell us that. But when we map in next year's projected data, things jump around-a lot.

Then we lay in next year's projected revenue, and all the bubbles jump.

Now we've got five variables in play: name of company, platform, features, revenue last year, and revenue next year. Before we add in more (and we're going to), let's see what we can see. First, the merged SMS-Peridocs surpa.s.ses us in revenue (bigger bubble), and their combined software surpa.s.ses us in features (their bubble moves up). At the same time, their merger will force them to combine two proprietary platforms, making their platform even less open than before (their bubble moves left). Meanwhile, our revenues have grown slightly (slightly bigger bubble), our continual software tweaks nudge us up a bit in features (our bubble b.u.mps up), and, a.s.suming we go through with planned platform Band-Aids, we are slightly more open (our bubble nudges right).

The post merger SMS- Peridocs surpa.s.ses us in revenue and features but becomes an even more proprietary (closed) system, while we marginally increase features and slightly open up our platform.

Meanwhile, let's look at what has happened on the open standards side of the plot. All the sudden revenue increases and feature upgrades of the old guard don't look so impressive. By the end of next year, it's projected that Univerce will not only exceed our revenues, they'll also beat us in number of features. How is that possible?

Next year the growth of the old guard pales by comparison to Univerce and MoneyFree, the new arrivals. Suddenly they've got more features and revenue growth than we've ever experienced.

In order to see what's going on, we need to plot in yet another layer of data. But before we do, we're going to need to make some room. Let's erase some of the details we've acc.u.mulated so far and pick things up by recalling the software improvements that Jason was demanding from us: flexibility, security, and reliability. In the past, proprietary platforms like ours were more secure and reliable than open platforms, although less flexible. To show that on our plot, we can just divide last year's landscape right down the middle: more secure and reliable on the old guard side (left); more flexible on the new arrivals side (right).

In years past, proprietary platforms were inherently more secure and reliable, while open platforms were generally more flexible.

This is why any Band-Aid increase in flexibility on our platform will decrease security and reliability: We'd be moving our bubble to the right without taking the security/ reliability line with us. But over the next couple years, it's expected that open platforms will improve so much that they'll become as secure and reliable as our systems are today-and remain more flexible as well. In other words, the companies with systems built on open platforms are not only going to offer more flexibility, they'll be able to offer as much security and reliablity as those of us with closed systems-if not more so.

The whole landscape is going to shift next year as open platforms improve. They'll offer security and reliability equal to (if not better than) our closed platform without losing any of their greater flexibility.

We can finally see what's really going on in our industry. As early as next year, the new arrivals-companies that came late and built their systems on open standards-are going to be able to offer services equal to or better than those of us who started early on with our own closed platforms. Which finally brings us back to our original question: Why spend $9 million on building a new open platform when we could spend a lot less on more moderate improvements to the platform we already have?

Believe it or not, we've now collected everything we need to show why. We started this chapter with a simple question: Does knowing anything more about our customers tell us why sales are flat? Using the six fundamental frameworks of visual thinking, we've not only answered that question (yes-we're not pleasing Jason), we've seen exactly how to go about keeping our customers happy (improve security, reliability, and flexibility) and stay the leader in our industry (move to an open platform). The problem is that it's going to cost $9 million. Which means there remains one more thing to do: Share these pictures with our executives and get them to see the same things we did-to see why for themselves.

In the next and final part of this book, we're going to walk through a short executive presentation built around nothing more than the pictures we've just created. In doing so, we'll answer the two remaining "big" questions about visual thinking-those that I am asked every time I talk about solving problems with pictures. First, what's the best way to effectively show a picture? Second, does a good problem-solving picture always have to be self-explanatory?

CHAPTER 15.

EVERYTHING I KNOW ABOUT BUSINESS I LEARNED IN SHOW-AND-TELL.

There are two remaining big questions about visual thinking, tough questions I'm asked every time I talk about solving problems with pictures. Both relate to selling ideas with pictures, the time when we need to finally share with somebody else the pictures we've created. The first question has to do with us as presenters: How can we best go about verbally describing a picture? The second has to do with our pictures themselves: Are they "bad" if they require any explanation at all?

Everything I Know about Business I Learned in Show-and-Tell Walk into a kindergarten cla.s.s and (with the teacher's permission) ask for a show of hands on how many of the six-year-olds can sing. Every hand will go up in the air. How many can dance? Every hand. How many can draw? Every hand. Now ask how many can read: a couple hands might rise. Then walk into a tenth-grade cla.s.sroom and ask the sixteen-year-olds the same questions: How many can sing? One or two hands. How many can dance? A few. How many can draw? A couple. Now ask how many can read. Every hand will go up.

Don't get me wrong: There's certainly nothing wrong with learning to read. But what happened to singing, dancing, and drawing? Once we believed that we knew how to do those things-in fact, at kindergarten age most of us practiced them happily every day-so why, ten years later, do so many of us forget what we once knew? And by forgetting (or even just thinking we've forgotten), are we missing something fundamental in our innate problem-solving abilities that could be useful to us in the black-and-white, right-and-wrong, quant.i.tative world of business?

As we reach the end of this book, I have one final story to share, and it's the best example ever of how not to present a problem-solving picture. It's a scary story and on the surface may appear to undermine much of what we've talked about here-at least that's what I thought when it took place. Only on reflection did I come to realize that the story, in fact, makes the case for visual thinking stronger, especially since addressing it forced me to go back and look at my approach to visual thinking all over again.

A year ago, I was hired to join a team of business consultants working on a huge technology project sales pitch. Each member of this team was handpicked for his or her proven expertise in a particular field, and each had been all over the world selling and leading successful projects. As I stepped into the conference room to meet them for the first time, I was already impressed. If you planned to spend $100 million on a new enterprise-wide technology system, these were the people you wanted: They just looked right.

Although I was brought in to help out just on the charts, I had a wonderful time working with this team, and even succeeded in convincing them to use pictures during key parts of their sales presentation instead of the usual bullet points. Having seen audiences fall asleep after the second page of bullets, the team was all for it, and after nearly three weeks of work we were all amazed by what we'd been able to accomplish. Together we'd managed to boil down a hundred pages of material into just six handouts and a dozen slides, without compromising any of the core materials and without losing the overall storyline of the proposal.

The showpiece of the presentation was a multiple-variable plot similar to the one we just created for SAX Inc. It ill.u.s.trated the client's industry by mapping together several variables (compet.i.tors, market share, industry work flow, sales over time) that were individually familiar to the audience but had never before been seen together in one place. The result was a picture that offered up numerous insights. It showed that the client's business model placed them at several unconnected steps across their industry; it showed that while they led in two of those steps, they lagged in others; it showed that their biggest compet.i.tors focused on dominating only single steps, etc. In other words, it was a picture that could launch any of several fascinating conversations, all of which were important to the client's decision-making process and all of which the team was prepared to run with.

As the chart guy, I didn't have a speaking part on pitch day, so I was given the unfamiliar role of sitting in the back of the auditorium where I could judge audience response and take notes for debriefing later. When our team entered the auditorium to deliver the pitch, I was ready to be amazed. I was, but not for the reasons I'd expected.

Lauren, the team leader, opened the pitch brilliantly. She was a great speaker-charming, engaging, loud. She led with a funny anecdote that got a chuckle from the room full of client executives, technologists, and finance people. It couldn't have been a better start.

But then she hit the "next slide" b.u.t.ton, looked up at the multiple-variable plot with its four layers of seamlessly integrated visual information, precognitive attributes, intuitive coordinate system... and froze.

It was like watching a cartoon: Lauren's mouth opened but nothing came out; her eyes darted across the fifteen-foot projection screen but saw nothing. As Lauren stood there, hands locked in midgesture, the room held its breath, waiting for her to explain what they were looking at, what it meant, and why they should care. But no sound was heard. I twisted in my seat, agonizing, barely able to keep from shouting out, "Lauren! Just say what this chart shows and start pointing!"

Mercifully, I managed to remain silent, and Lauren-the consummate consulting professional-wasn't going to let a bunch of colored bubbles on a chart knock her off track for long. She took a breath, recovered her composure, and said, "We created this chart to show where you sit in your industry. Next slide please."

We didn't win the project.

In the debrief we all agreed on what had happened: Although Lauren and the team now knew how to create a problem-solving picture, we'd never discussed how to talk about one. When she got up on stage in presentation mode, Lauren's mind expected the slides behind her to contain words in lists, something that she'd spoken to hundreds of times. But when she turned around and saw colored b.a.l.l.s and bits of text connected by lines and arrows, her mind went blank. Where was she supposed to start? What was she supposed to say? Other than the headline and the labels on the coordinates, there was nothing there to read: no bullet points, no summary, no words.

I knew at that moment I'd stumbled upon the greatest challenge to solving problems with pictures: Although we know how to look, to see, to imagine, and to show, n.o.body since kindergarten has told us how to talk about what we see. Just like singing, dancing, and drawing, we once knew how to show and tell, and we did it without bulleted lists. Not anymore.

For a time, I despaired: Was there no future for anything other than simple tables, Venn diagrams, and bar charts as presentation tools? How could that be, after all my research and personal experience in seeing how well pictures worked? Then I remembered the English breakfast and the countless other pictures I'd worked on with teams across dozens of companies in half a dozen countries, the pitches I'd seen won based on nothing more than a single chart that the CEO immediately "got," and the project teams that understood what they were supposed to do only when they'd reviewed that detailed Gantt chart. No, I thought, the problem isn't with the pictures-the problem is in remembering that show and tell are two different words.

Then it hit me: We already have the answer, and just like the visual thinking process itself, the answer is something we all do all the time without even being aware of it. In fact, the process for talking about a picture is the visual thinking process. Let me show you what I mean. Let's go back to SAX Inc. for a moment, and make that final $9 million pitch to the executives.

Look, See, Imagine, Show: The Four Steps of Selling an Idea with a Picture Quick review: We've created a series of pictures to help us solve the problem of flat sales at our accounting software company, SAX Inc. Those pictures lead us to a possible solution-spend $9 million to completely rebuild our software platform. OK, that's one problem solved, another created. How are we going to convince our executives to spend $9 million on a major project when we have flat sales? To address that, we created another set of pictures. We mapped our executives' decision-making process, exposing cause and effect with a flowchart so we could see what we'd need to show, and then we prepared an elaborate, quant.i.tative, visionary, comparative, forward-thinking picture to tell them the whole story.

Imagine that we've scheduled a meeting to present our ideas to the execs. We're in the conference room thirty minutes early, preparing for the execs to arrive. No worries. The way we're going to approach this is exactly the same way we made our pictures: We are going to take the execs with us through the four-step visual thinking process as we look at a landscape of information, see those things in it that matter most, imagine what they mean, and then show the result. The only difference is this time the information landscape is a plot we've already created, and we already know exactly what we want to show.

Look, see, imagine, show. We've done it before and now we'll do it again.

As we're waiting for the execs, we're not booting up our computers, looking for wireless connectivity, or trying to hook up the projector that never shows the right resolution, but that doesn't mean we don't have pictures to show. And we're not stacking up color-printed decks in front of each seat, but that doesn't mean we don't have sheets and data to hand out at the appropriate time. No, what we're doing is drawing our picture on the whiteboard, as big as we can. We're sketching in the coordinates and first four variables of our plot (compet.i.tors, platform, features, last year's revenue), preparing to convince our execs why by engaging them in an interactive (truly back and forth), live (but that doesn't mean unscripted), back-to-basics (but that doesn't mean simplistic) visual thinking session.

This is what we draw on the whiteboard before the execs come into the room-the t.i.tle, coordinates, key, and first five variables of the plot.

With our drawing done, we sit down and take a breath. Right on time, the execs arrive. Our execs don't like small talk these days, so we stand up and immediately direct their attention to the whiteboard.

"As we all know, we've got a major problem to solve. Sales of SAM have flattened, and if we don't get sales back up in the next year, we stand to lose our top spot in the market. Our group believes that we've identified a solution, and we want to share it with you by taking you through this visual overview of our market."

Brief aside. The fact that we've got an elaborate picture drawn up on the whiteboard is already working in our favor. Since the executives can immediately see that we've got something well thought out in mind, but can't completely understand everything on it, they are anxious to hear what we have to say. They'll likely even give us a moment more than usual to get to the point. This is when we start looking aloud.

Look: What's the picture all about? What's included and what's not? What are the coordinates and dimensions?

Looking aloud means that we aren't going to toss our executives into the middle of the metaphorical bowling alley. We're going to take their hands and walk them there, pointing out the coordinates and dimensions of the place as we go, giving them a moment to figure out where we are and what we're supposed to do now that we're here.

With that approach in mind, we start the tour of our picture. "Our goal in creating this model was to build a baseline of our industry according to several critical factors, ranging from platform to feature set to revenue. We believed that by looking at the business in this integrated way we would see our problem in a new light, potentially illuminating new and unexpected approaches to solving it.

"There's a lot included here-and there's going to be a lot more-so let me quickly show you what we have. First, we looked at what types of compet.i.tors we face, whether running on proprietary or open systems, which we plotted here along the bottom." We point out the horizontal axis.

"Next, we asked what kind of features each company's software offers, whether a full suite or just a few. We plotted that here, going up the side." We point out the vertical axis.

"Then we added in last year's revenues using proportionally scaled bubbles plotted onto the appropriate quadrants of the chart. You see us up here in the lead with revenues last year of $25 million and the fullest feature set running on our proprietary platform, while you see MoneyFree way down here, with few features running on an open platform, and next to no revenue." We point out the bubbles at the extreme ends of the scales.

We look at the execs and see nods; they're with us so far. Time to let go of their hands and take a step back: We're about to drop a bomb.

See: What are the three most important things that stand out? How do they interact? Is there a pattern emerging? Is there anything critical that we don't see?

Seeing is about pointing out what's most important in the picture-something that we haven't even drawn in. So, as we say, "Here are those same companys' revenues projected for next year," we draw in next year's bubbles starting with our own quadrant, explaining about the SMSoft-Peridocs merger, etc., then draw in MoneyFree, and finally Univerce.

One by one we draw in next year's bubbles, starting in our corner and saving Univerce for last.

"Not only is Univerce expected to grow ten times in revenue, it could very likely surpa.s.s us in features as well, knocking us into third place in offerings and size." Boom.

Our executives see the point now and the questions start to fly. Some are defensive, like, "That can't be right. Where did you get those numbers?" Some are aggresive, like "What in the heck is Univerce up to?" Some are cautiously exploratory, like "Hmm-is there anything we can do?"

The first question we answer precisely because we know exactly where the numbers came from, and that's when we hand out the detailed data spreadsheets we created while researching the picture. The second question we answer by describing next year's antic.i.p.ated increase in security and reliability on the open platform and the immediate impact that it will have on sales of open software. As for the third question-"What can we do?"-we're ready for that one, too. "Thank you for the perfect segue," we respond, "let us take you through two possible options that we've identified."