As a Product Manager, most of the time you will be in a position to influence others; your boss, engineers, designers, and clients. Thus, communicating with data will come in handy in most situation.
Last updated
Was this helpful?
I believe that most of us have some experience in dealing with data presentation in our professional life. Even back in university, me and friends had joined several business case competitions that required compelling data presentation skills, not because we already mastered the skills, but we need to learn from our peers, regardless. Of course, I think thesis defence was such good occasion that some, if not most, of us had experience on how to truly convey a sound argument backed with highly relevant and valid data.
Based on my poorly done observation, I noticed that while many of people I know are good with data or information analysis, they often struggle to communicate the findings and its insights. I admitted that it also happened to me quite often just the same 😔. What I usually found is, and sometime did myself sorry ✌🏽, delivering presentation with a graph or chart slide after slide without clear context and correlation between slides that made the audiences felt overwhelmed with incoming flood of information. I bet you are already familiar with this kind of situation.
I often ask myself, "How could I be much better at giving data presentation, just like those seasoned public speaker?"
Because as a Product Manager, communication is your main job. I would say that you are partly hired to communicate very well, at all level. And communicating with data, or at least with an educated-guess, is your main weapon to win everyone else's attention.
In her book, Cole fleshed out the step-by-step she frequently uses to conduct SWD. I will shortly re-brief those steps here before we go with real-life example scenario that may happen in our day-to-day life as Product Manager.
First thing to remember is that, when it comes to presenting with data, we are to do explanatory analysis. So, let's take some times to turn the data into information that can be easily digested by an audience. Although we are often tempted to show everything we did on our exploratory analysis, it's not necessary the right move for you to walk the audiences through that process (again). Because it's your job as the communicator, not them.
Next, for communication in general, there are at least three questions to answer before doing anything to help us setting up the right context; Who, What, and How. These are what mentioned by Cole about those three questions:
To whom are you communicating? It is important to have a good understanding of who your audience is and how they perceive you. This can help you to identify common ground that will help you ensure they hear your message.
What do you want your audience to know or do? You should be clear how you want your audience to act and take into account how you will communicate to them and the overall tone that you want to set for your communication.
It’s only after you can concisely answer these first two questions that you’re ready to move forward with the third: How can you use data to help make your point?
The pilot summer learning program was successful at improving students’ perceptions of science and, because of this success, we recommend continuing to offer it going forward; please approve our budget for this program.
Clutter is visual elements that exists inside the display space, but do not add any meaningful value to support the main idea. This instead will add cognitive load burden to the audiences, so that they need to process those meaningless elements in the middle of understanding the actual message. Therefore, it's necessary for us to eliminate the clutters to ease out the audiences' minds, as Cole said:
There is a simple reason we should aim to reduce clutter: because it makes our visuals appear more complicated than necessary.
The steps to remove clutter is called "Decluttering", and it's mainly dependent to Data Chart you have available in hand. But the general guides are: a. remove chart border, b. remove gridlines, c. remove data markers, d. clean-up axis labels, e. label data directly, and f. leverage consistent colour.
When we're talking about visualisation, it couldn't be separated from the domain of design. There are at least three design concepts which are very relevant to communicating with data: affordance, accessibility, and aesthetic.
Here's the case for you:
You're invited to a year-end meeting with upper management and one topic that was being discussed is "Why did the product growth get slowing down since early Q4, which was shown by the plateau of Weekly Active Users curve?". Many arguments came up during the meeting, but one of the solution that almost everyone agreed upon was "Simply to acquire more users to a point where more users will convert to Active Users, eventually!".
While that might be an obvious answer for everyone, your gut feeling told you that there's a missing link here. But, since you didn't have enough information to convey the argument, you decided to gather more data first and then craft a sound reasoning shortly after, to influence other stakeholders regarding the thing that you believe as the true problem.
You knew that Active Users is basically the sum formula of new users plus returning users. So, after the meeting is done, you immediately asked one of your data peers to query and flesh out the data. And here's the new chart you got from your colleague:
From the chart above, you shall notice that most of the active users are driven by new users. And we do also notice that we have fewer returning users by Q4. I would then assume two key findings from this information alone:
We're currently heavily relying on Users Acquisition to drive the active users metric, therefore it's rather unlikely that acquiring more new users to our product will bring sustainable growth, long-term (in terms of active users).
While there's a growing base of returning users from Q1 to Q3, we see fewer people come back in Q4. It could be a strong sign that users do not find enough value to continue using our product, or that something hinders/prevents them to discover that value. Therefore, spending resources to growth in this stage is rather an unwise move.
As the PM, you need to influence your stakeholders to reconsider the next course of action to resolve the seemingly growth issue. You want to invite them to take a step back and redefine the problem statement, because based on your rough analysis it might be not a growth issue at all, but is actually more on product-market fit problem.
, or SWD for short, is the book title written by Cole NK. As the name suggests, it provides the idea on how to tell a "business" story using data. So instead of just showing a bunch of data and let audiences making sense out of it themselves, Cole argued that we could actually presenting a storyline built on top of those available data, to guide the audiences arrived at the conclusion we intended them to buy in.
Additionally, by the time you have figured out the main insight from the analysis you want to convey to the audiences, you must be able to convert it into one short full sentence that wrap up the whole . This will help you to stay focus and not easily tempted presenting any findings that not necessarily relevant with the main message you want people to hear. Here's the example from SWD book:
There's a lot of chart or graph types you can choose to visualise the data. Each type will work best for certain dataset and context you want to display. For example, Line chart is typically used to show changes across timeline, while Bar chart is usually used to show categorical data. I recommend you to visit the beautifully made by SWD team, listing all the chart types they frequently use in their works, with great length explanation of each!
While it's true that the previous decluttering will help audiences have easier time to read the data, we can further make life easier for everyone by drawing their attentions to particular data points we want they to focus on, to help them making sense out of the data faster. And we can achieve this result by using preattentive attributes, a visual cue that can be easily scanned by people's eyes. It basically means, something that is more standout than the rest will draw our attention first, the moment we take a look at it. There are several preattentive attributes we can leverage for our data visualisation, and you can find more about them here: .
