Analytics has become an increasingly competitive market, and it’s no wonder why. This type of platform is a goldmine for important insights that businesses can leverage to make better strategic decisions, which makes purchasing this type of software a critical investment.
But, seriously, how useful is a tool if it’s severely lacking in the UX department?
Too often, turning the plentiful data from the analytics software into useful insights is a burden that falls squarely on the shoulders of the user. Much to the customer’s disappointment (as well as to my growing frustration), useful, easily-gleaned insights are nonexistent within the tool itself.
As technology continues to advance, customers are demanding a better user experience of all software companies, especially analytics tools. The bar has been set fairly low at this point, leaving plenty of room for innovative UX design and functionality improvements. Users don’t need data for its own sake; they need to answer questions, come to conclusions and discover what works versus what does not work.
I promise this is not just a rant. I’m actually going to talk about what analytics software creators can do to give customers what they want, by solving for common UX challenges that are inherent in many analytics products. A better UX will ultimately increase the value of your software, boost adoption and retention – as well as cure a couple headaches of my own.
Showing Data Without Context
Imagine if your boss told you, “I rate your work quality as a 5.” You would have no idea if that is a good thing or a bad thing. Is it 5 out of 5, 5 out of 10, 5 out of 100? Is 1 the best score or would a higher number be the best? If your boss said, “I rate your work quality as a 5 out of 6, with 6 being absolutely perfect,” then you would know what it meant.
In the same way, displaying metrics without any context only confuses users and hinders the effectiveness of the software – it forces your user to do extra work to understand whether that number is better or worse than before. Unfortunately, analytics tools today are sorely lacking qualitative analyses that help the user understand the data and read between the lines.
The ability to understand what the data is saying is paramount. Without this context, the software provides only numbers while the user is searching for understanding off of which to base next steps.
Solution: If a user cannot gain insight from a single data point in isolation, then as a rule, analytics software should always show data in context, either through a comparison against a previous time period, a ratio against one or more other metrics or showing its progress toward a goal. You can also enhance clarity of interpretation by including plain-english explanations via text or graphs.
The future of analytics software is the ability to provide predictive and prescriptive analytics that display where the foreseeable trends are heading, as well as suggest specific actions as to what should be done to help the user reach their goals.
A stunning example of a company contextualizing data to maximize objectives is Netflix. They looked at pause, rewind, and stop analytics for one of its most popular shows, House of Cards, and then used this data to craft the plotline and character twists. While it’s impossible for Netflix to know exactly why viewers pause or rewind during an episode, they can ask and assume. Their analytics team put these events in context and the results are later used to improve the viewing experience.
That is what analytics programs need to start doing: providing context and actively assisting the user to understand the implications of that data.
Distributing Metrics Across Separate Screens
I cannot wait for the day when businesses no longer have to use Microsoft Excel sheets for reporting (guilty). Here’s a hint: if your users are creating spreadsheets to do reporting, your analytics software is missing something. Too often, the information of analytics tools forces users to jump from one report to another on different screens, making the reports extremely difficult to compare or look at as a whole. Remember our rule: one data point alone is not enough to draw any kind of useful conclusion – so siloing metrics into separate screens isn’t going to help the problem!
Solution: Allow users to overlay metrics on top of each other in one graph, to increase the likelihood that they’ll spot a meaningful correlation between metrics.
Sorting through analytics takes so much time. Why add to that? Presenting meaningful metrics together on one screen will greatly enhance the usability of the software, while also increasing the user’s’ happiness with the software.
Here’s the other thing to consider, too: Ease of use on multiple screen sizes is extremely important to the success of a dashboard. Depending on the data being looked at, some industries may require it in real-time and on the go, which means — you guessed it — mobile interfaces for analytics programs. What does the mobile user really need to see? What use cases would require, or even benefit, them to look at metrics on a smaller screen?
Part of a great user experience is anticipating the user’s needs. This is definitely one of those situations. It means you are required to really know your user before designing. Don’t worry, research helps with that.
Consistency is key with UX in analytics tools, because it brings clarity and predictability to the user. Inconsistent or irrelevant labels will only make an analytics platform more challenging to use and understand.
For example, do all of your users know the difference between a Session and Visit? What about industry acronyms such as CPL, CPU, COGS, etc.? If your users do not understand what every metric and report means, they’ll never get to the insights they truly crave.
Solution: Utilize consistent labels, but use them how people expect them to be used. Encourage easy data visualization by utilizing the most common language and actions that users are accustomed to when choosing labels and displaying charts.
Being overly creative and displaying information in unintuitive layouts/labels makes it hard to digest information and also increases the likelihood of mistakes.
Misuse of Color & Data Visualization
Many analytics dashboards rely on data visualization to turn the data into a story that users can easily comprehend. In order to effectively achieve this, charts must utilize different colors and visual metaphors in an organized, sensible manner.
For instance, you’d expect to see positive numbers in green (or at least black) and negative numbers in red, right? Innovation in visualizing data & metrics is great, but stay focused on solving the interpretation problem for the user. No one will care about your gorgeous multi-colored bubble chart, if they can’t understand what it’s telling them. Resist the urge to get cute with colors and contradict what a user would expect to see.
Additionally, charts that do not offer enough variation in colors are difficult to read for many people and can be impossible to read for the millions of people who suffer from color blindness — and that definitely hurts usability.
Solution: Craft a color palette that has a wide range of hue and brightness, because these types of colors are easier to depict from each other. Use natural gradients that are commonly seen outside and are comfortably familiar to people when used in design.
And stay consistent! Make sure that your color language is tightly defined from the get-go (e.g. “Green = Better, Red = Worse”), and do not deviate from it – if necessary, expand your color dictionary, but don’t mix and match aesthetic colors with others that are meant to convey concrete info to the user, it’s just a recipe for user confusion.
Consider creating palettes with at least eight colors to efficiently display complex data in a chart. (Here are some of our examples.)
Report Loading Speed
People are filled with urgency, especially in a business environment. Users want tons of accurate data immediately and waiting for even a minute could seem like a lifetime. In a world where most information can be accessed with a quick press of a finger, you can’t blame people for their impatience.
Slow-loading reports only exacerbate the problem of analytics research being an incredible time sink. By engineering your reporting queries to be executed quickly, and for the screens themselves to be quick-loading, you’re doing your users a huge favor, and reducing their hesitation to commit the time to pull up your reports in the first place.
Solution: Do everything possible to minimize waiting time for users, which goes without saying.
However, If loading times cannot be reduced, then use a progress indicator to minimize anxiety or frustration for the user. Utilizing a progress indicator for any wait time of more than two seconds will increase user happiness and confidence when accessing the platform. Plus, it gives the user an idea of how long they need to wait will help suppress the impulse to jump ship.
Creative loading animations are increasingly popular; they give the users something pleasant to look at while they wait and reassure them that the software is functioning. Some software designers use loading bars, spinning wheels with the percentage of completion, and some even create tongue-in-cheek scrolls to keep the user entertained — things like, “crunching the numbers, replacing calculator batteries, calling grandma to say hi,” you get the picture.
These can be a lot of fun and provide some moments of delight for the user while they wait.
Proactively Notify Users of Abnormal Patterns
The goal of analytics tools should be to get the user to that “AHA! I’ve learned something useful I can act on” moment. The design should be focused on drawing users’ attention only to the places where there’s potential gold – i.e. abnormal performance, either good or bad.
Personally, I believe that logging into analytics to check that everything is okay should never need to happen. That is why analytics is installed in the first place, to tell you when something significant has happened.
Imagine if your alarm clock rang every hour — even though you didn’t need it. Now imagine your alarm clock only rang when you actually checked the time. Both are useless. Just like you set an alarm for a specific time when you need it, your analytics should be watching for a specific condition you set, so you can forget about what the time is until you need to be alerted.
Solution: The better way to serve users relevant and timely data is to auto-create dashboards based on the user’s goals that will allow them to see their most important reports in a clear, concise way.
When something is abnormal, analytics tools should reach out and get the user’s attention. The software should be smart enough to learn over time what’s normal and what’s abnormal, and notify the user by their method of choice (email, SMS, Slack, push notification, etc.) when something out of the ordinary happens with their key metrics.
Help Users Help Themselves To Data
Analytics platforms are meant to help users uncover insights regarding what happened. But someday (hopefully soon) they will also make educated predictions about what will happen if trends continue and prescriptive data that suggests which changes should be made. The platform should present data in a clear, consistent, customizable and familiar format, as well as give the user the ability to present information on one page (or screen).
The more steps you make a user go through to complete an action, and the longer you make a user wait for information, the more frustration builds up and decreases the likelihood continued use. By using these tips to improve the UX of your analytics tools, you will increase the platform’s value and establish an advantage over your competitors.