Over the past few years, I have made a good number of reports and taught a number of BI classes. Besides the technical aspects of how to build reports, I always focused on the usability and credibility of the reports. Usability and credibility are essential if you want your reports to be used more than once by the audience you designed it for. And if someone does not trust your reports the first time, he will probably not come back a second time to see if you improved your report (that is if you received feedback about what was wrong). You typically only have one chance to impress, so the question is: how to make credible reports?
In order to make credible reports, I think you need to:
- Understand the business requirements
- Find the right data
- Use the best possible visualization (e.g. there is such a thing as a pie chart with too many slices)
- Apply a consistent styling
This blogpost is all about this fourth step in the report creation process as for the first 3, I hope it is obvious why they are important. A consistent style is often overlooked as it requires some extra design effort but is important for the following reasons:
- Reports look more professional
- It takes away the need for interpretation by the end user as the consistent style makes the way your report looks and feels predictable.
This is why I defined for myself a set of BI design guidelines that I wanted to share mine as an example but do not forget to define your own before you start creating your reports:
- Choose a standard layout
- Add Introduction Text
- Add No Results Layout
- Add Conditional Formatting
- Add Totals & Performance Tiles
- Make reports actionable
- Add Navigation Paths
Here are some more details and examples of each of them and why I think they are important.
1) Choose a standard layout
As stated above, defining your standard styling guidelines is important as it makes your reports look more professional and it takes away some of the interpretation that inevitably is happening if a report is not clear about what information it is trying to share with the end user.
Here are some of the things I try to keep in mind:
- All my report titles look the same and I use the standard styling options to make the report titles look like all screen titles in Sales Cloud.
- All my tables are formatted in the same way and I use the standard styling options in BI to make tables look like list screens in Sales Cloud.
- Choose colors carefully. Users tend to interpret colors according to their cultural background and as a report designer you need to be aware of that; Red indicates bad news in the western world but actually is the color of prosperity and good fortune in some Asian countries.
Maybe you should standardize on some other colors to indicate positive (orange) or negative information (purple). Choose whatever you want but use the colors consistently over all reports. In the example below, every user using the report below will for a moment wonder whether or not the red for the ‘High Technology’ industry, although it is the industry with the biggest pipeline, still is under performing to expectations just because the bar is colored red. Is that what the report designer intended? Nobody knows 🙂
- Every currency value (revenue, …) must have a currency code indication, every percentage value should have a percentage sign, …
- All performance tiles should look alike. Here is how I format mine:
- All my performance tiles show the value in font size 24 or 36 and in the same pale dark blue color
- All my performance tiles are 200 pixels wide and 80 pixels high and have the same dark blue color 1 pixel border
- All my performance tiles show the description of the value in small default color font below the value
- All my performance tiles compress the values
I know it requires some time to apply these to your reports, but I assure you this is time well spent. Your reports will be more trustworthy, will look more professional and will be used more often! There are ways to make it easier to apply consistent styling to your reports as I have detailed in my BI video tutorial recording on this topic
2) Add Introduction Text
Any effort to take away from the end user the need to interpret the results shown in a report is energy well spent. We are all used to see legends next to graphs, as they are great to explain what used colors in a graph stand for, but that is not all that we can do? Why not add some information to help the user understand your reports. This can be done by a simple little introduction text as shown at the bottom of the table below, or by adding some help icons that show a tooltip when hovering over them.
3) Add No Results Layout
Sometimes reports for good or bad reasons do not return any data. The default message end users get to see in such a situation is sometimes more confusing than usefull as a technical reason why no data was returned is being shown. It is very easy to replace this with a less technical, more meaningful message.
Adding a meaningful ‘No Results’ message is as simple as it can be. In the report editor, open analysis properties where a custom No Results message can be added.
4) Add Conditional Formatting
Adding conditional formatting to tables is a great way to make it easier for an end user to understand the data displayed in the table, especially in large tables with lots of columns and lots of rows. Conditional formatting is all about adding font colors, background colors, images and more to draw the eye to the information inside a big table that needs to be noticed.
In the example below, finding the years where closed revenue went up or down has been made a lot easier by adding the red and blue arrows no?
5) Add Totals & Performance Tiles
As stated above a few times already, tables are typically all about sharing a lot of details with end users. The aggregation level is typically much lower than what you will find in a typical pie or bar chart. But there is no reason not to help end users in better understanding a detailed table. The easiest way to do is by adding some type of totals to your reports.
By default, totals are shown at the bottom of a table. Especially when you have a table listing a lot of records, in order to see the totals, you might have to scroll down or even skip through a number of pages before you get to see them. I consistently add performance tiles above tables as a way to make grand totals easily accessible.
6) Make reports actionable
Most often, reports are created to analyze data and take action upon the findings. The easiest way to do so is by allow users to drill from a report into the application where changes can be made, and upon saving or cancelling out of the screens, the user is returned to the report.
I try to have every account, opportunity, lead, contact, … record name to be shown as a drilldown link, making the report actionable as he can directly from the report change data in the application or initiate task and appointment (or more) to follow up on his findings.
Here is a recording on how to configure your reports to make this happen.
7) Add Navigation Paths
Let me take the previous guideline a step further. Why would you present data in a table full of records from where an end user can drill down into the application in the first place?
It forces the end user to go look for the record he/she is looking for before a drilldown can happen. This can be hard in big table and from a design point of view, as it is hard to predict how many records will be shown in a table so I most often just assume there will be too many and a way to help the end user find the data he is looking for is needed.
Best practice should always be to add a way for end users to filter the data in a table in order to focus on just a part of the table results (to reduce the amount of result records in a table). Most often this is done through prompts, but a much more intuitive and visually attractive way to achieve this is available to; by drilling down from a graph into a table where the graph serves the purpose of a prompt.
In the example below, by clicking on one of countries, represented by the vertical bars in the bar chart, the table will only show customers for the chosen country. The bar charts serves as a prompt. I always try to only allow for end users to use tables after they have gone through at least 1 graph to filter the table results.
I really hope you understand how important all of this is, how with a little effort you can make your reports look a lot more credible and professional. Feel free to share your guidelines with me, I am always interested to learn how others implement their reporting strategies and guidelines.
