Thursday, February 26, 2015

How to create a monitoring and evaluation dashboard

So you have a website, a blog, the usual social media channels on Twitter/Facebook/YouTube, maybe a series of RSS feeds. On top of this, your organization or research programme also publishes original content, or indexes content produced by others into an online portal. And you also organize events and workshops and maybe offer grants and awards.

With all these online spaces, outputs and products that you produce, how are you going to collect and aggregate this data as part of your monitoring and evaluation activities? And how are you going to display and present it in an effective way that can be easily understood by your co-workers, managers and donors?

For the past couple of years, I’ve been experimenting with tools to display data and information in online dashboards. This post presents a short introduction to the topic. It’s the first of a series of posts that will look into online tools for data collection, storage and display.

What is a dashboard? 

According to Stephen Few’s definition “A dashboard is a visual display of the most important information needed to achieve one or more objectives; consolidated and arranged on a single screen so the information can be monitored at a glance.”

More generally, a dashboard is a visualization of data related to the operation and performance of an organization, program or service. Dashboards are helpful tools to provide you with a quick overview to see whether you’re on track to reach the objectives stated in your logframe or Theory of Change.

Note that information about a wide range of channels can crowd one screen. So it’s important to be flexible and keep users in mind - keeping scrolling *very* limited and using features like tabbed navigation to view different set of metrics and indicators.

What are the steps to follow to build a dashboard? 

The Idealware report Data at a Foundation's Fingertips: Creating and Building Dashboards presents an excellent and detailed step-by-step description of the process to design effective dashboards for non profit. Ultimately the process boils down to 4 main phases:

  1. Define your audience
    Of course this is absolutely critical, determining the way you design it, the graphs you include, their order and sequence. The dashboards I’ve developed in the past were mainly designed for managers and executives, to tell them about the progress of a program or service at a quick glance.
  2. Identify the metrics to display - and how you collect them
    With the tons of metrics that you could collect, and the space limitations of a dashboard it is important to agree upfront which ones will be displayed in the dashboard. So it requires a bit of negotiation to agree upon what’s in and what’s out. Of course, the metrics should be useful in terms of monitoring progress towards the objectives in your logframe and theory of change. In this phase it is also important to discuss collection methods, frequency and access. 
    • Are there any process that you can automate? 
    • What is only possible instead through manual data collection?
    • And is it realistic to collect this data monthly - if the properties are high traffic or include active campaigns, for example, - or is quarterly more realistic?
    • Where are you going to store the raw data and who should have access to it?
  3. Identify your dashboard platform
    This is a maturing market so there are a lot of possible solutions - from expensive business intelligence software to low cost or free tools. Generally the decision is defined by the resources available as well as the time you and your users have to invest in learning new tools. Note that while potentially you can build a dashboard in Excel, investing some time in learning how to use a powerful and flexible dashboarding tool such as Tableau Public can enable you to design more complete and effective dashboards.
  4. Sketch, prototype and roll out
    In the design of the dashboard you need to find a good balance between the amount of information you want to display and the limited space available. So you have to carefully decide what graphs and chart you will use, what explanatory text you should include, which colours to use when...This will take a lot of testing and iterating to find the optimal design. Bottom line, your final product should: 
    • Be simple, intuitive, easy to read and understand; 
    • Present data together from different sources in an uncluttered way and following a logical sequence or order; 
    • Offer a quick overview of the key metrics and indicators to assess progress towards the objectives of your program/organization/service. 
In the following posts, I’ll be presenting two case studies about work we did recently on visualizing monitoring and evaluation data into online, interactive dashboards. I will look specifically at the tools  used to put these dashboards together, as well as the individual tools used to collect and store individual indicators and metrics. 

For the more techie readers, I’ll also share the details of what I’ve learned recently using Google Apps script to automate some data collection and storage processes, as well as tips and tricks to monitor activities and engagement around Twitter, which I’ve been experimenting a lot with lately. So stay tuned!