The Perfect Dashboard Rocks
By Till Buettner of Deutsche Post DHL Group
Thanks Brian, and hello everyone! It’s great to speak to such and audience at the Virtual Analytics Summit. And I’m glad I can present to you the perfect dashboard or how it can rock your work.
So, let me start and you know this I think have seen before. If not, especially this one, but you have seen dashboards like this one. And as it is perhaps nice to get drunk and have some analytics therapy session with others, most time you want to have or to achieve value with our day to day work.
So let me give you some examples. These are typical examples I have seen before. And it doesn’t matter from which tool. This is from Adobe Analytics; but the same happens in Google Analytics or other tools and you see reporting where you have perhaps some tables, a few graphs, a line graph, or bar chart. But it doesn’t tell you anything. Perhaps you have also just reporting with tables or with nothing in. So It doesn’t help anyone at all to see this.
A short summary that I would like to talk today about is:
- I want you, after this, to have a clue about how to understand your stakeholders. And it’s one of the important things you can do. You should do.
- How to define a structure and design which makes a dashboard perfect.
- And at least customize the content so the dashboard is right for your stakeholders. Because as we all know, it is not the one and only perfect dashboard out there. So no one can claim it for himself or herself and say “I have the perfect dashboard. I rocked it!” No. We have all done this, at one point in time thought, ”I’ve done the perfect dashboard. It’s so good,” and in the end the stakeholders look at it and don’t understand it. So, we want to have a look at how we could be defined, structure, and make the perfect dashboard for your stakeholders.
As we want to make the perfect dashboard, I want to go through the word “P.E.R.F.E.C.T.” It starts with “P” for people. Talk to your stakeholders and ask for their business goals. And if you do so, really talk to them and listen to them. They always have quests and you should go in to this quest. And ask the right questions for them. It’s about goals and requirements from them. Don’t talk about metrics or KPIs or stuff like this. Most times they want to do this. Most times they are like “Yes, I need page views.” Don’t do this! I normally say, “Sorry but the 90’s called, they want their page views back. And why do you want to have page views in your dashboards? Let’s talk about this.”
And then, as an example, it turns out that they want to know how many people read the articles. Then you can go through and say, “Wonderful, that’s a requirement. So let’s talk about how we can define how an article is read. Per time, per percentage, scrolled on the page, or whatever. But that’s a requirement.” So please, if you talk to your stakeholders, get their requirements and their goals.
The next letter is “E” in perfect, and it stand for empathy. So, you need empathy to understand your stakeholders, and to put yourself in the shoes of them. How you can achieve this is perhaps in using the empathy map. I don’t know if you’ve heard about it before, but the empathy map is about getting yourself into the view of your stakeholders. Get what they do, how they felt, what they thought, and what they said. Most times, the things they said is not what they thought and really not what they felt. It is about getting the clue about what they really want. Sometimes it is like, “yes, we need this xyz in the dashboard.” And then you ask and get into and ask requirements and stuff. In the end you get that that isn’t really what they want. They had just heard or got the pleasure to have this in their dashboard, but it is perhaps not what they want to work on. So, listen to them, ask and do active listening. Repeat what they told you. Reflect how they behave to you. If you do so, you will hopefully get back the idea of what they really want to see in the dashboard.
The next letter is “R”, it stands for relevance. So, focus on the business success and break down by relevant dimensions.
So, what I mean with this, is if you have a dashboard, make clear. Get the relevance for the stakeholders. Get a visualization which matters. This one is perhaps not the best one because it uses red or green to show good or bad. But perhaps your stakeholder has problems with red or green so perhaps you have to take out the colors. If you do so, look at your stakeholders and get the relevance for them in the dashboard.
One important thing is, if you talk about format, again ask the business what suits them best. So what I mean with this is, we as analysts, are aware of using most interfaces, doesn’t matter if it’s web or whatever. Most times we like to use the interface so we are using Adobe Analytics or Google Analytics in the dashboards in there. Perhaps, the stakeholders can’t use the dashboards in there or perhaps it’s overwhelming for them.
So, perhaps, it’s better to take all of the data without an API into Excel and then present it into PowerPoint. We all know that PowerPoint is out of hell, but most of our stakeholders are into PowerPoint. They learned how to read PowerPoint, how to work PowerPoint. So, if it’s the format for them, then do it because you do the format for them and not you.
If you do this for stakeholders which are also analysts or IT guys or whoever then perhaps it’s the raw data that is good to work on. Then work with so many things out there that you can use to do dashboards in a good way and to have special needs accomplished.
Focus on the solution that fits best for the stakeholders, and it doesn’t matter who is the stakeholders. So, if you talk about stakeholders, it is not the only man on the level, it could also be IT, it could be another analyst. It doesn’t matter, it just matters who you’re working for in this moment.
Next letter is “E”, and it stands for elegance. If you talk about elegance please carefully choose the structure for your visualization. So, if we say carefully choose, you could over harm a stakeholder with what you show them in a dashboard. Too many numbers, to big numbers, no numbers, just graphs, whatever you take in there. It could be distracting for them. So, carefully choose the structure.
And as you can see in here, it is 2-20-200.
If we have a dashboard like this in here, we’re talking about a 2-20-200 rule which I got from Marco Hassler, a German analyst, doing a great job, written a book. In here, he talked about building dashboards that you can in two seconds see what is good or bad. In these two seconds you can directly get if it’s going up or down or bad or good, as I said before. In 20 seconds, you get the clue about what is really good or what is bad. So, you have the possibility to think about the things you’ve seen before. In 200 seconds, you have the possibility to act. If you see that there is something bad, and you have the possibility to think about what is bad, then you can go deep into the information and then act. As you can see in here, it doesn’t matter if it’s top down, if it’s mixed, it just matters that the structure is defined and told to the stakeholders and that they learn to handle it. If this structure, as you can see here, fits best then that’s wonderful. If the stakeholder needs the two seconds at the top then the 20 down then the 200 in the bottom, that’s fine. However you want to structure it.
The next letter is “C” for context. In here, dashboards should tell a story. It doesn’t matter if you take PowerPoint or Keynote or whatever to do the dashboard. It is not about telling a story like a book or something like this. It is more that people understand what from where to where you want to lead them in the dashboard.
Let me give you an example. In here you see the dashboard. In the right hand you see that the 2-20-200 rule is kind of integrated. So, in the middle you can see in two seconds how good or bad the data is. In 20 seconds in the top, in 200 seconds down. It’s really interesting in here, we have a story about attract, engage, convert. This story is not for me, it is done by Urs Boller as the whole presentation here is done by him and me together. So, thanks to Urs for doing this presentation more than once with me. Today, I’m doing this alone, but all of the ideas in here are done by us both.
What you can see in here is that he has built a story about how the website attracted the users, and by which campaigns. How the people engaged on the website. In the end, how many people converted. So you have a short story, and it could be a much longer, a bigger story. It begin, should fit to the needs of your customers and stakeholders.
The last letter is “T” and it stands for text. People love explanations. In here, I will try to keep that a little more into. What do I mean with “people love explanations”? Most people don’t like to read data in a table. They read data out of somewhere. We, as analysts, have learned that the best thing we can do is transfer data into visualizations. So we do. Often, dashboards are from top to bottom built with many many visualizations. Bar charts, graphs, scattergrams, whatever. Also this is perhaps for stakeholder overwhelming because they didn’t learn how to read the visualization. If you want to get the stakeholders, you should first get your head around who is in front of you. If you’re doing this to IT guys or data guys, perhaps just tables are the best things because they want raw data. If you have advanced stakeholders, c level, or management level who are into visualization and have learned to read them then go for visualizations. But in the end, every one of us has learned is to read text. Every one of us has learned to read a magazine, a book, or whatever where something is written down. The idea behind this, why don’t we transfer data into text?
So, in analytics, it is possible to change the name of metrics into text. It doesn’t matter, you could also get your data into Excel let's say, and then change the data into text or change it in a PowerPoint. So, if we talk about, let’s say, you have 5 percent uplift in your conversions last week, and from 1 to 5 percent, it’s an average uplift. Why don’t you say that it is an average uplift? Why do we need 5 percent? Because most stakeholders don’t get a clue about how much five percent is. Is it good? Is it bad? It is average. Wonderful, it is average. So why don’t we write down, “We had an average uplift in the conversions last week.” So, if you do stand ups with small dashboards for stakeholders then perhaps just two, three, four sentences with just words and no data is the right way to go. Like, “We had an average uplift in the conversions last week. The most people by far came by tag bookmark traffic, and the bounce rate goes down to perfect normalization level.” Something like this.
If you do so, stakeholders hopefully will interact with the data more. If they get the taste of what you presented, then they perhaps deep dive into the data. Then the dashboard with visualizations or raw data is perfect to deep dive into. To get the taste for them, the text is perhaps the right way to go.
On the other hand, each dashboard should have a glossary and a contact. If you present data to stakeholders, have the most interesting metrics or KPIs written down so everyone is aware of what you’re telling them. If you just write conversions as the metric, and conversions could be anything, it could be the newsletter signup, it could be the transaction. So have a glossary in there where you define what metrics or what KPI is how defined.
And the other thing, it’s much more important than the glossary, is the contact. Why is the contact so important? On one hand it is important because we could write our contact in. So, if there’s a problem with the dashboard or they want new things in, they can reach directly to us. They don’t have to find someone or if you’re working in a bigger team then you have a special person. A special email address where they send these to. They can have it sent to.
On the other hand, it’s very important to have the business owner in the contact. Because you’re doing this dashboard for someone. Someone has told you that they want this dashboard. So, if there is a business owner, get his email address and name or telephone number in there so people can ask the business owner, “what is the dashboard about?” or “what are the requirements?” If you’re talking about the dashboard and if the goals are reached or not, they can talk to the business owner about the right thing.
And you, if you look at the dashboards, have all the information to who you can connect if the dashboard is really needed or not. You don’t have this information somewhere else. So, the contact is about ownership and who is responsible for these dashboards.
Short summary, again, understand your stakeholders. Get yourself into the shoes of the stakeholders. Use empathy to know or to get knowledge about what they really want and what they need in the dashboards.
Define a structure and design. If you define this, think about 2-20-200 rules, and think about the format, how to deliver the dashboard. Which is why, if its Excel, if it’s PowerPoint, whatever you use, just use the right format for your stakeholders.
And in the end, customize the content because you could have distance to the stakeholders perfectly if perfect structure, designed very well, everything is wonderful, but the content is everytime the same so the stakeholders wouldn’t work with it. So, each dashboard, is individual piece. So, if you have done one and two, good. And perhaps you’ve defined also the template which could fit most times for structure and design, the content should always be customized for the stakeholders.
So, in the end, “It is not the layout which makes a dashboard PERFECT. It’s how it helps business to get fast insights and leads to the right actions.” That’s what it’s about, and a last sentence on all of this. As Urs and I did this presentation the first time, we thought about having slides, giving them out to the people. Most time slides, again getting on drive and hard disk will never be seen again. What we did is a small website about this. This perfectDashboard.rocks, and the idea behind it is to have a community which talks about dashboards, what ideas they have if they build the dashboards. It doesn’t matter which tools you use. Most content on the website in the moment is Adobe Analytics related, but that’s just because we are working with Adobe Analytics. If you have dashboards in Google, in Tableau, and whatever you’re working with, and you want to share and to talk about this, get out there. We have a block we try to promote each dashboard and help each other. There’s no business behind here. This is just knowledge transfer for our community. I would love to get in touch with you there more.
So, there are the scenes. It is Urs doing much of the work and me a little bit of doing some presentations. It was wonderful to do this presentation for all of you, thanks for listening. If there’s any questions feel free to ask.