You might have heard multiple times, that beautiful dashboards are attractive, memorable and people want to spend time with them. But more often, it is enough for your dashboard just to look professional. So, the readers will be sure that this one is not made by an amateur who does not know what is he doing, and who does not even know how to handle data and make calculations.

This “unprofessional” look typically comes from the creator being still in his experimentation phase (just look at my first data visualisations) and suddenly discovering that you can add a gradient and 3D effect to a chart on Excel. That might look cool for the first time, but for the first time only –by seeing a chart with gradient and 3D, the reader might think – oh they were making their first chart.

Also, the “unprofessional” look might come from the lack of understanding of basic data visualisation principles. The creator might just drop some charts with default settings, having no idea why to even bother changing those.

Quite a lot of teaching materials online provide ways how to create a WOW effect, create a stunning dashboard, YouTube is full of such tutorials. Usually, the goal is not a dashboard providing a natural flow of experience, but a STUNNING one. To stun seems to be the goal. In business settings, the natural habitat of dashboards, they are viewed regularly, so you don’t need to stun your readers every Monday morning when they just want to see how bad the previous week was. The dashboard just needs to be professional and trustworthy because wow and stun effects wear off quickly. (Unless you’re a consultant – so you just make a stunning dashboard, run away, and never return.)

If the dashboard looks like made by a professional, there will be fewer doubts about the trustworthiness of the data, and this effect will stay there for a bit longer. So let’s explore what might make a dashboard to look like one made by a professional.

Align strictly

Nothing is more ridiculed than the Comic Sans font. While it comes with tremendous benefits in readability and accessibility, it is simply not aligned well, so it looks childish, messy, and untidy. The same goes with a dashboard, where boxes, titles, images are not aligned – this simply does not look like made by someone who cares about what they do.

Just align everything carefully. Tableau aligns things by default with containers, meanwhile Power BI provides guidelines which show which items are aligned with which other items.

Reduce colours

A children’s colouring book sparkling with all primary colours having saturation cranked up to 11 is an antithesis of the dull corporate world where everyone wears a light blue shirt and dark blue suit.

You have to reduce colours to look professional, and by reducing colour we mean two things: reducing the number of different hues and reducing the assault the colours make against our eyes. Make them less bright, less fluorescent, desaturate, choose pastel colours. In the end, we will have a colour palette consisting of two to three main colours, each of them having a few additional shades, and that’s it.

Square corners

While square corners look cute, cuteness is not something needed for a professional looking dashboard. Meanwhile, squared corners communicate strictness and precision. We can have shapes in charts to have squared corners (bars, markers in a line chart) or other objects (boxes, buttons, icons). There are cases when it is OK to have a more rounded and bubbly design, but make sure this aligns well with the branding of the company you’re making a dashboard for.

In the end, most shapes in charts are squared by default, so the point is not to waste effort on introducing roundness. If you want some shapes rounded a bit – just consider how much of a radius you want for them – if it’s too much, the whole thing will look like a game interface rather than a dashboard.

Conventional layout

Experimenting with layouts may lead to a unique and, in some cases, much better experience than the traditional title-navigation-numbers-charts-filters-on-the-left approach. But think for a while about financial reporting – the format has not changed for at least hundred of years, any introducing novelty would mean that the creator is simply unfamiliar with the field. I as a finance professional myself, have seen countless fancy dashboards trying to display financial information in a “better” way, and they all seem ridiculous to me: “Those people definitely have no understanding of finance!”

If a reader considers the author of the dashboard incompetent in a particular field the dashboard is about, it might hurt the author’s feelings, but more important it will hurt the trust of the numbers. Read The Big Book Of Dashboards and do as in their examples.

Simple, straightforward charts

When I first discovered my passion for data visualisation, I created a chart for a presentation which defied all the rules in this article and my boss told me: “Well, the partners will ask whether we’re doing business or f—ing art. But let’s keep it.”

The problem arises if you create a chart which makes business people feel that they don’t understand. They will not feel dumb themselves, they will think you are dumb and giving them nonsense.

That’s why it is a very safe idea to stick with the basics – bars, lines, dots to display most things. Business people feel quite well if pie charts are around, but some of them might have already heard that PIE CHARTS ARE BAD, so they might frown on you if you introduce one. Just stick with the basics.

Legible fonts

Fonts have a tremendous impact on the overall look and feel of the dashboard. Once again, let’s mock Comic Sans for being childish – obviously this font will not be perceived very will in a dashboard, however, no one even tries to do that (I hope?). But legibility is another topic – make sure that the fonts you use are clear, readable and not too small.

The King of France, Louis XIV, ordered a creation of a geometrically strict and balanced font Romain du Roi to convey the superiority of royal documents compared to handwritten ones or those printed in not so perfectly shaped letters. If a king cares about fonts, you should too.

Do not show errors or (Blanks)

If there is a weird default message somewhere in the dashboard, it looks like there is an error in the dashboard. Usually, it is an user overfiltering data, but they get distressed when they see (Blank). “The data is missing!” — scream they with despair.

To avoid this rapid deterioration in trust of your dashboard, simply make sure that whenever there is a change to see (Blank) value or any kind of default error, add some safeguarding. The most obvious one is IFERROR function in Excel, which prevents error to showing up. For Tableau or Power BI, there are similar functions to avoiding blanks or errors.

Check all actions

Check actions in Tableau, check bookmarks in Power BI. While those things are powerful ways to create a great User Experience, with great power comes great opportunity to seriously screw things up. If some action or bookmarks return weird results, do not work properly, anyone using such a dashboard will simply say: “OK, this thing is not working, I cannot use it”

That’s it. I had a case when a bookmark was returning weird results on one chart in one Power BI page, so users simply skipped that page because it was “broken”.

Make sure there are no hiccups in interactions by putting your cat on your keyboard and let them emulate user behaviour. Or, for better effect – give your dashboard to a real user and see what they do.

Think about everything in advance

I have heard this phrase many times: “Oh, you have thought about this in advance, nice!”. People really appreciate that. If you don’t think about some edge case, some force majeure that might happen, or some incredible creativity of users on all the crazy ways how to mess up with filters, this is acceptable. But if you have thought about it, then it is a bonus. It means you are a true professional because you probably have experience and you immediately know where the issue might be.

Because you screwed up with that previously.

So, spend at least a few moments in thinking about what could go wrong.

Performance is key

We are slowly drifting away from pure Data Visualisation topics and arriving at the field of User Experience, which means Everything.

What does poor performance mean, how it is perceived? Think about a laptop which cannot handle a new game – isn’t that laptop a piece of crap? So if the dashboard is slow, it will look like a piece of crap. Furthermore, I have experienced when a junior developer just loads a table (I’m just loading a table, what could go wrong?) and all the other reports break because computation capacity is exhausted. Poor performance is often a sign that a report created by someone who does not know their BI tool really well.

Performance optimization will be unique in every architecture, and it is a field that senior developers care more about. So, care about it.

Proofread

If you miss a typo, haven’t you missed any rows in the data? This is a ridiculous line of thougth (sic), but seeing typos will seriously undernime (sic) the trust in the author.

Take some time to proofread all the text in the dashboard once again. I know we are dedicating time to numerous things now already. This is where those long hours of finance professionals are coming from – they have to deliver a perfect and polished presentation by tomorrow. Such things take time. Or drugs if you don’t have time. Proofread.

Definitions

Every so often, when creating a new dashboard, I am not sure what one or the other indicator means. Then I typically just expect that readers with domain knowledge will know themselves what means what. This is a sign that I might not know what I am doing.

By providing definitions, you are not only signalling that you understand what are you doing, but also reassuring the reader that what they think some ratio means is exactly what it means.

However, if you define EBITDA for financial audience, they will think that you’re probably encountered this term for the first time, that is why you included it in definitions. So, they cannot trust you because you don’t know what you’re doing. The best is to consult key users and ask what definitions they would like to see in a report.

Information about sources

For infographics – including sources is a must, but for dashboards that is often skipped. Usually, the creator of the dashboard himself is not aware of the exact primary source of the data – they’re just pulling it from some SQL database. And frequently the source is obvious – of course you’re getting your financial data from ERP.

However, if the source is not obvious, it might be a good idea to include an indication where the data is coming from. What is absolutely necessary is an indication of how fresh the data is, especially for dashboards that are viewed daily. Do the numbers I am seeing account for the full week, or just part of the week? Is yesterday included? Is this morning included? People will trust you more if you just include an indicator on when the data is updated or what is the latest date point.

Of course, if the report is not updated for too long, the effect will be the opposite, the dashboard will look abandoned and obsolete. However, rightfully so. Some data governance procedures would prevent this from happening.

***

We learned a few points what we should do, to make the dashboard look like made by a professional. There are also a few pitfalls where we can overdo, so let’s watch for these!

Bar charts only

Don’t overdo simplification of charts. Lines, lollipops, in some cases even treemaps or pies are useful. Forcing one solution for every problem means you are not choosing the best solution for every problem. Unthoughtful use of bar charts only will make your dashboard seem limited, made by someone who is not wielding the full power of the tool they’re using.

Everything in one brand colour or blue

Don’t overdo simplification of colours. It is awful to see a dashboard where all bars look like all buttons, which look like all filters because they all are squared and have the same colour. Probably no one will judge you for this, or lose the trust in the dashboard when there is just one colour. But colours used right will enhance the user experience rather than confuse. If done right. (Here is an article on colour)

It is OK to colour bars in other, not the main brand colour. The main brand colour might not even be the best one. Use secondary colours – most brands have them, or some variations of the main colour.

You might already know that colours have associative meanings in human psychology, and that blue symbolizes responsibility and is suitable “for corporate sites or designs where strength and reliability are important“. Don’t just do everything in blue just because of that. Unless you work at a bank, then probably do everything in blue.

Countless tables

Tables provide precise numbers and may seem very trustworthy, but in some cases this backfires. First, the performance of tables is often not great because the report has to pull many numbers at once. Second, it suffers from all the troubles which data visualisation is there to remedy – it is slow to read, difficult to understand and is ugly. “It was not thought through in advance very well!” Finally, you might end up with scroll bars, which is a terrible User Experience for a dashboard.

Reports with countless big tables usually are made by less data-savvy self-service users, who just want to see numbers, but have no idea how to create proper data hierarchy and navigation. It is just a data dump and not a proper dashboard, and this does not look professional.

And sometimes the requester of the dashboard is a less data-savvy self-service user, who just wants to see numbers, and has no idea how to create proper data hierarchy and navigation. They just need a data dump to export to Excel. I’m not talking about such cases here. That’s a topic for another one about dealing with requests.

Images

Images are controversial, and they might be considered chart-junk. But if they are illustrating the point, do not feature people or cats (those attention stealers!), it might even add to the appeal of the dashboard. Business people are just big children in the end.

Sacrifice UX

Never ever sacrifice User Experience for any decision that you make for a dashboard. If the specifics require you to add more colours or use an unconventional layout – add them. If some additional calculation will dramatically reduce the performance of the dashboard – find a workaround. If the Definitions section takes too much space – move it to another page or make it a popup.

In the end, our goal is not to make dashboards look like they were made for a bank or insurance company, the goal is to invoke trust. The point of design is not beautification of the report, but guiding the reader, and letting them know that the report was made by someone who understands what they are doing and cares about the reader.