Building a Professional Business Intelligence Platform on a Non-Profit Budget

Big Data’s importance in driving revenue and success at large, for-profit companies is well documented, and has been an integral part of successful business strategy for years. However, with the often staggering price tags associated with data (infrastructure, human capital, etc.), many nonprofits may feel priced out, and unable to compete. Thankfully, there are some ways for nonprofits to take control of their data, without breaking the bank.

Data Collection

There are a number of analytics tools that will help you track your web and mobile traffic. For its ease of use, versatility, and industry-standard metrics, I prefer Google Analytics, powered by Google Tag Manager. Tag Manager allows you to easily customize your event, dimension, and metric deployment — a huge benefit for companies with limited web development resources. There are also a number of pre-written tags that you can use: we have based much of our custom tagging work on Simo Ahava’s blog. If you have some money to spend, I’d recommend Google Analytics 360, over the free version. At about $150,000/year, 360 is a considerable investment for a non-profit, but its benefits are tremendous for those who want deeper analysis, and who can afford it. But don’t worry — I’ll cover cheaper options as well.

I prefer Excel to Google Sheets. That said, Sheets has at least one advantage: the Google Analytics plugin. Do you need to constantly update a spreadsheet with numbers from GA? This is the answer to your prayers, and it’s free! You can even publish the sheet as a .xlsx file, and use it as a data source for reports you need to run.

Data Analysis

Now that you’ve stored that data, how do you make sense of it? Google Analytics offers a pretty good out of the box solution for making sense of your traffic, but more in-depth analysis may require some additional firepower. Here’s where GA360 comes in. If you’ve made that investment, it includes the ability to analyze your raw data through unsampled reports and Google BigQuery, a powerful database that unlocks a lot of additional analytics flexibility. Data is loaded automatically from Google Analytics to BigQuery, and the setup is easy.

BigQuery gives a generous 1 Terabyte of free queries per month, and a $300 credit towards your first year of use. Organizations with smaller datasets may not even use the whole $300, but YMMV. BigQuery connects nicely with a number of Business Intelligence platforms, and can answer just about any business question you may have.

If you’re not able to spend six figures on premium analytics, there are workarounds. Google’s API Explorer gives you much of the same flexibility as BigQuery, but it’s free (with limitations). Connect to Google’s APIs through Python or PHP script, and you can load data to a data warehouse on your own.

Data Visualization

Looks tasty, but not the most compelling data viz.

Most of us don’t naturally see data materialize into insights the way Neo does in The Matrix. Strong visualizations are critical for pretty much anyone trying to comprehend large data sets. That’s where data viz software comes in — and no, Excel probably won’t cut it.

I’ve been using Microsoft Power BI for several years, and I’ve been impressed with its balance of power, ease-of-use, and cost. For the most basic dashboard and report creation, Power BI is free, but you’ll have to pay to unlock some of its more useful features, such as sharing data with co-workers. Fortunately, Microsoft offers a $3 per user monthly option for non-profits (you’ll have to go through their simple and fast vetting process). If you’re not terribly concerned with data security, and you’re comfortable sharing as embedded iFrames or non-authenticated Web URLs, you can forego the paid version. However, if you need more security for your data sharing, you’ll need a Pro license — for every user who creates or accesses reports. At $3/month, however, it’s still a reasonable charge, especially compared to the $9.99/month our for-profit friends shell out. Unfortunately, Power BI is only available for Windows at this time, so if you’re a Mac user, you may want to take a look at other options.

Power BI connects to basically any data source, from .xlsx or .csv files, to Web APIs, and just about any database you can think of. BigQuery connectivity requires a free ODBC driver, but once installed, it’s easy to pull in multiple SQL queries to power your reporting. This connector also allows you to pull data from BigQuery directly into Excel. There are native connectors for Google Analytics (best used for very simple data sets), Facebook, and more. One of the advantages of Power BI is its automation — most data sets using cloud-based sources can be set to refresh on a schedule, without having to manually update anything. Who just refreshed 15 reports from Disneyland? This guy! You’ll need an always-on PC to schedule the refresh; processing power isn’t terribly critical, but it’s helpful to have a good internet connection.

For Mac users, I’d recommend Google Data Studio. I prefer Power BI’s almost unlimited range of connectors, but Data Studio is also a robust and easy-to-use visualization platform. It will work best for organizations who store the majority of their data in Google services, MySQL or PostgreSQL DBs, but .csvs are also supported. For a small fee, a number of companies have developed Data Studio connectors to other data sources as well.

Challenges Yet to Come

You’ve implemented the technology, you’ve built some reports, and you feel good about your progress. As well you should — but you’re not done. It’s critical to make sure you position these reports and dashboards as indispensable tools in your non-profit’s mission. In any organization, for-profit or non-profit, you’ll probably encounter some resistance to data. As my closing thoughts, I offer a few best practices for winning the hearts and minds of your co-workers.

• Design for the user, not for yourself. This is tough, even for BI veterans. We all come into a new job with our own biases and preconceptions about what’s important, and what people’s focus should be. Solicit feedback from your stakeholders, run beta tests, take note of any points of confusion or frustration, and make any changes you need. BI, just like software development, should be iterative, and you should be open to your users’ concerns and questions.
• K.I.S.S. Bigger isn’t always better, and less is more. Strive for simplicity in your reporting, and make sure the most critical KPIs are the ones you’re highlighting most prominently in your reports. The flip side of Power BI and Data Studio’s great visualizations is the tendency to overuse graphics, and to overwhelm your audience. Do you really need that fourth scatter plot? How does it tell a different story than the first three?
• Plan ahead. Power BI allows you to use a single data set to feed multiple reports — this is a huge advantage, if you approach it the right way. Take some time to plot out your data sources, the metrics and dimensions you’ll need, and the types of reports your organization will use. Building a solid master data set can take some time, but you’ll be able to quickly spin up an almost unlimited number of reports that update every time you update the master. Automatically.
• Have Fun! If you’re working with data, chances are you find it interesting. So why present a bunch of Excel tables in Ben Stein’s voice? Data is exciting. Don’t put your audience to sleep. The way to make sure your audience is engaged is to make sure that you are engaged.