Recently, I’ve been frustrated with some of the articles I’ve read about building great analytics dashboards. Their titles often led me to believe that they contained must-have insights about creating dashboards that engage users, but in the end, most are about use of color, pretty charts, and the latest feature a particular vendor has launched. And in my experience as a data product builder, these things aren’t what will make your embedded analytics product a success.
In the past, I followed the similar rules for dashboard design to those I found in the articles and yet, we often didn’t generate enough user engagement to persuade the customer to continue paying for analytics. I built dashboards with cutting edge animations, that displayed data in the best possible format, and which looked modern and frankly, beautiful. And I’ve seen many of them fail to engage users in the long term. As I’ve discovered since those early days, the rules of dashboards are different when applied to analytics-based products. "Pretty" simply isn’t enough.
“If you build it, they will come…”
A great line in a movie, but it often doesn’t hold true, at least for data products. For an analytics product team, it’s not uncommon to “build it” only to have users fail to walk out of that cornfield or to have the users that do arrive head back into the corn after just a few visits. When you’re the person responsible for bringing a product successfully to market, that’s a big problem. Luckily, through the judicious application of mistakes and failed attempts on my own products, I’ve learned a few lessons that help increase data product adoption both in the short-term and over the long haul.
Through trial and error here’s what I found: I, like many product owners, was targeting the wrong users with the wrong functionality. I’d often go after the “core user”—the person sitting in front of whatever application into which my analytics would be embedded—thinking this would have the most impact. When adding analytics to a CRM system, I would think about analytics for the salesperson. Order management system? I might focus on analytics for the order entry clerk. While it seemed logical, it failed more often than it succeeded. As I learned the hard way, when you launch your product you need user engagement, you need it fast, and you need the right kind of user. Enter the CEO.
Why the CEO persona?
Now, when I start designing the initial analytics dashboards for a product I build the CEO dashboards first. Of course, I get to other personas later, but I start with the CEO. Why? Because in addition to being the person that ultimately controls the project budget, the CEO is the person that sets the decision-making “style” for the entire company. Get the CEO hooked on using analytics to drive decisions and the only way that she’ll give up such a powerful tool is if you pry it away. Solve her problems and she’ll be your champion and drive adoption throughout the business. Sit in on an executive staff meeting where the CEO knows more than their subordinates about an issue because of dashboards and watch how fast adoption spreads downward.
Get inside the mind of the CEO
If the success of your data product depends on generating engagement and the CEO is the most effective place to begin to creating that engagement, how do you get started?
In previous articles I’ve described a methodology for picking the best analytics for your user personas. You pick a persona, identify their “mission” or work goals, map the workflow they follow and any gaps that exist. The analytics you build should directly correspond to the gaps you’ve found in the workflows for each persona.
This is a great process for ensuring that every analytic placed on a dashboard is tied to the needs of a persona, but it doesn’t work well for the CEO. A CEO has such broad responsibilities, such varied “missions” that creating workflows and finding gaps/pain points is a difficult proposition. So I use another technique: empathy.
You begin the process by understanding the traits common to many top executives and then building dashboards that work with, rather than against those characteristics. Here are a couple to consider:
Characteristic #1: Extreme time pressure: Most CEO’s are busy people. Between customer visits, tackling the problem of the day, interviews, and board meetings, it’s amazing that they have any time at all to review operational performance. Like me, you’ve probably had countless meetings scheduled with top executives that get pushed out days or weeks due to various crises or obligations. Time is at a premium for the CEO and the limited time they do have is vitally important. You can use this to your benefit.
Characteristic #2: Anxiety: Not the crippling kind of "can't get out of bed" anxiety, but rather the “I have a lot on my plate and none of it is trivial” type of stress. A high level of anxiety is common for CEOs and with good reason. If they make mistakes, they can risk the entire business. People can get laid off, stakeholders can lose money, businesses can fail. A lot is on the line. Additionally, in a larger business the CEO can’t possibly understand all of the key issues at play across all of the many organizations. Due to the scope of the role, she’s operating with incomplete information. You’d be a little anxious too. But just as a little anxiety can be a good thing when so much is at stake, it can also be a valuable tool for the analytics product leader when designing dashboards.
This is the world of the CEO, whether building customer-facing SaaS applications, running a manufacturing business, or leading a non-profit. With an understanding of a couple of the pressures that they face, you can start to design dashboards that will reduce their anxiety and time concerns while producing high engagement.
Three elements of a great CEO dashboard
A word of warning: What follows below might not be what you’re expecting. It’s distinctly lacking in information about aesthetics, color theory, or optimal chart selection. Though these elements are important, none are key to generating CEO engagement (and hence, revenue).
Here’s what is key:
Design for Triage
Imagine a scenario in which you’ve come across the scene of a accident and people have been hurt. Injured victims lie across the ground in front of you and they need help now. Where do you start? For emergency responders, there’s a protocol known as the “ABC’s” in place. You examine each each of the injured for Airway, Breathing, and Circulation. Depending on the results, the responder may treat the patient immediately or move on to a more critical situation. This is done in part to relieve some of the extreme time pressure on the emergency responder—take too long, treat the less critical patients first, and people can die. The CEO, also under time pressure, has a similar need to identify the most critical “patients”, take action, and move on as soon as possible. You can generate CEO engagement by designing your dashboards for easy executive triage.
Triage based dashboards should to do the following:
- Show an overview of the situation
- Identify the most critical “patients”
- Indicate how badly each is “injured”
With the limited time available, the CEO user must be able to open the dashboard, receive a quick synopsis of key operations via key performance indicators, and understand which areas need immediate attention to ensure the health of the business. Unfortunately, many dashboards aren’t designed with CEO engagement in mind and require clicking, scrolling, drilling, and interpretation to see the big picture. For the rare occasion when the CEO has the time to spend exploring the analytics, this is fine. For the every day situation where the CEO needs to gain an understanding of performance and move on to the next task, it makes the dashboard frustrating. It’s easier to ask a subordinate to perform some analysis and report back. The time and effort required to triage business health with your dashboard is too high. There goes your engagement… If you want to create an engaging dashboard for the CEO, make sure that it’s designed for easy problem triage with the most critical issues front and center.
Design for Relationships
The CEO role is fundamentally different from the jobs most of us hold. We might have responsibility for a team, a product, or for a quota, but in most cases we have limited scope. We can focus our attention on that single team, region, or department. The CEO job is different.
Imagine if you—the manger of an order processing team—had the following schedule:
- 8 am: Review yesterday’s performance
- 9 am: Meet with irate customers who have delayed orders
- 9:30 am: Conference call with analysts and press to discuss the future of order processing in the 21st century
- 11 am: Review space planning with the building facilitates to ensure room for future growth
- 12 pm: Lunch with key investors in order to reassure them that their investment is safe
You get the idea. If you had a schedule like this, you’d have a tough time catching your breath, much less achieving an understanding of the strategic needs of the order processing team. From hour to hour, the CEO can be working on highly disparate projects, all of which are important, but for which the common thread tying the activities together might be hard to see. Engaging dashboards will help the CEO to see these patterns.
Your CEO analytic dashboards must show not just “snapshots” of performance, but a coordinated picture of the interrelationships between key elements of the business. In past years, this might have been displayed as a “balanced scorecard” of metrics from finance, operations, customer satisfaction, and employee satisfaction. In recent years with all of the amazing analytical tools at our disposal, we’ve gotten away from the “balanced” approach and now favor dashboards based on data readily available and which can be displayed in a attention-grabbing format. These look nice, but don’t help the CEO look across the business and therefore, don’t offer an incentive to keep using the dashboards. A big picture dashboard showing cross-organizational relationships does. Here’s what you need to do:
- Map out (at a high level) the key processes of the business such as “develop new products,” “sell products,” “install products,” “support customers,” etc.
- Break each key process into 3-5 phases—identify the metrics for each of those phases. These metrics summarize health within the process, but aren’t so deep down as to bog down the CEO.
- For each key process, create a high-level set of 2-3 summary metrics that captures the overall process health.
- Combine the summary metrics and the phase metrics into a single “process health” block for your dashboard.
- Arrange the executive dashboard by these “process health” blocks in order so the CEO can assess health from development, to sales, to deliver, and so on.
Combined in such a manner, the CEO can view the performance of the business in a logical manner. This enables the understanding of relationships: “it appears that the selling process is going well, but look—we’ve got backlogs over there in the installation area. We’re going to start getting support calls soon. Better focus on that!” Showing individual metrics forces the CEO to mentally build the interconnected-ness themselves. Showing the relationships allows them to manage the business. And that generates engagement.
Design for Action
Triage and understanding of relationships is only valuable if you can do something about the problems you uncover.
Years ago, my team built a set of dashboards for executives to show performance of projects around the company. It worked well—over time we were able to shift operational reviews from PowerPoint-based meetings to web-based review calls. But I missed the “take action” part of the equation. There was no tracking of identified issues or tasks designed to bring resolution within our system. Each review of the dashboards would result in an executive assigning actions—verbally—to a manager who would dutifully note the task and then work the solution. Further follow-up was then completely offline, taking the form of emailed status reports or hallway conversations. Not a great way to generate repeat engagement for your data product… Big miss on my part.
When building a data product that will be engaging for the CEO, you need to allow them to take action right then and there. This might be in the form of an “Assign Task” button right next to a chart on the dashboard or it might be a simple message thread below the chart. What it shouldn’t be is an external email link or any other mechanism that forces the CEO out of the system to check future status of the problem.
Design your dashboard so the CEO sees the issue, requests an action be taken from within the analytic system, and then returns to the application to learn what progress has been made. Having the dashboard serve as the sole source to check status is a technique I call “loading the itch.” It simulates the CEO to return to your data product to learn what—if anything—has been done to solve the problem that she assigned. Like the drive you feel to keep checking your social media account to see responses and “likes”, the drive for the CEO to return to the analytics will be powerful and the result will be consistent executive engagement.
Don’t get me wrong, I really enjoy reading all of those articles about building great dashboards. I love learning about color, gestalt theory, and interaction design. But these things aren’t enough for the product team seeking to create a data product that generates revenue. You need to design analytics that get key decision makers—like the CEO— seeking out your dashboards on a daily basis. Understand the CEO, her key motivators and how you can help her better understand the performance of her business, and you’ve got an engaged user and likely, a new advocate.