12 Higher Ed Analytics Techniques, Tips or Tricks

You can slice and dice digital analytics data in many ways to discover useful insights.

Some techniques, tips and even tricks can make the difference between merely reporting data and informing the decision making process.

That’s why I asked the 12 higher ed speakers of the 2020 Higher Ed Analytics Conference to share something they’ve found particulary helpful when analyzing data leading to impactful recommendations.

Use custom dimensions in GA – Mandee Englert (Penn State University)

What I’ve found the most helpful is to add Custom Dimensions to our Analytics properties.

They have been extremely helpful in content analysis and journey mapping where we are now able to discover things like:

1. which group of news stories perform best on our site
2. how often we should be posting new content
3. which content is attracting our internal/external audiences
4. which content is helping to drive conversions like donations or visit registrations, etc.

Look at what happened just before and after – Tatjana Salcedo (Uni of Vermont)

To gain deeper insight on the page level, I like to look at which of our pages brought visitors there and where those visitors went after.

Two “tricks” I use are the “previous page” secondary dimension on the page of interest and, for the next page I create a segment with “previous page” equal to my targeted page and apply that to my pages report.

Come up with a smart formula – Corynn Myers (University of Michigan)

We apply a specific formula during our brainstorming sessions for recruitment campaigns that is incredibly helpful in estimating what it’s going to take to reach a specific goal.

Once we’ve gone through the formula process, we can show our partners “the world” and then have a realistic conversation about budget and app numbers. Want the formula? I will share it in my session at the 2020 Higher Ed Analytics Conference :-).

Embrace simple solutions – Avinash Tripathi (Kaplan Higher Education)

Keep it simple! Solving a business problem using a difficult or complex methodology necessarily does not mean it is correct.

Given times we use excel forecasting modules to derive to similar outcomes vs building a sophisticated propensity model using SAS, R or python.

Don’t overthink it – Emily Gustafson (Cornell University)

This isn’t so much of a trick but a tip – don’t overthink it.

There is nothing wrong with easy math as long as it gets the job done. If something seems too complicated, that means it’s too complicated. There’s going to be an easier way to accomplish your goals, and simple tools are so effective. I’m in Excel every day – no shame.

Alternatively, if your technology is proving to be a barrier to gleaning actionable insights from your data, there’s going to be something out there that fits your budget and skill level to get the job done.

Keep results simple – Dr. Nicholas Ladany (USD)

When possible, keeping the results simple to understand for multiple audiences.

Sometimes people working with data analytics provide results that are “over-sophisticated,” meaning that they do not translate well, nor do they read the audience at hand. It’s like a math professor who doesn’t reach students.

Don’t filter at the data tracking stage – Orion Stavre (WPI)

Always tracking the data at the purest form, and then aggregate as needed for reporting.

I know it sounds simple, but most presentations and training that digital analysts make suggest that the tracking is set up in a filtered aggregated form. I have found it very helpful to track data that mirrors elements of the site in a granular form, which allows for very robust tracking without the heavy (time consuming) management. On the reporting side, as long as you have a good basis on the tracked elements, it is still easy to aggregate the data and compile user friendly dashboards.

Segment your data to analyze it – Elicia Dennis (University of Notre Dame)

Custom segments in Google Analytics are currently the most helpful trick I have in my day-to-day data use.

By filtering out local traffic to remove web activity from faculty, staff, and current students, I have gained insights into the web behavior of prospective graduate and undergraduate students. To determine how our websites are being used during the admissions process, I focus on key time periods in the student recruitment and discernment processes such as the weeks leading up to application and decision deadlines. To get a sense of our current student’s web behavior, I have a custom segment that only shows traffic from campus IP addresses from users in the 18-24 year old demographic. These segments have given me a better understanding of the web needs of two of our target audiences.

Go visual to help stakeholder see results – Kris Hardy (Messiah College)

Creating reports and presentations that help stakeholders visualize the data has made a huge difference. Using scatter plots, treemaps, ribbon charts (among others) along with heat and scroll maps can really help bring the data to life.

I use Power BI (a tool by Microsoft) to develop these visuals. It’s only $9.99 a month and is pretty easy to use.

Unleash Google Data Studio’s power – Laura Montgomery (The New School)

This is more of a tool than a technique or trick, but over the last year and a half I’ve finally been able to make more fulsome use of Google’s Data Studio.

It’s an amazing tool for pulling in marketing analytics from multiple sources (Google Analytics, Google Ads, as well as plug-ins that pull data from social media advertising accounts) and creating dashboards for various audiences. I have a complex, multi-page dashboard I use for my immediate team, but also have a number of simplified dashboards I share with leadership and other institutional stakeholders.

Adopt Python to crunch big data sets – Aaron Baker (Harvard University)

I’ve been learning the Python programming language and in particular the Pandas data analysis library has been such a time saver for number crunching, especially with large datasets.

I was trying to use Excel to make pivot tables, but the data was too big and Excel would just choke. Now with a simple Python script using the Pandas library, I can get reliable results in a fraction of the time. I’m also experimenting with using Python to gather data from APIs and do the number crunching directly. I see a huge potential for this work.

Remember statistical significance – Joshua Dodson (VisionPoint Marketing)

This isn’t a specific trick, but what I find to be most helpful is to rely on actual statistics to make sure that what I’m saying is accurate.

It is easy to “measure” things and conjecture that one thing is performing better than something else, but have you run a chi-squared analysis to make sure that it is statistically better? It is important to take analysis to the next level to make sure that we can back our claims — not just with data that “looks” good, but with data that is verifiably accurate.

A conference focusing on higher ed analytics?

The 2020 Higher Ed Analytics Conference (#HEA20) is a must-attend event for higher ed marketing professionals and teams looking for inspiration, ideas and best practices to step up their analytics and measurement game in 2020

Read below what your higher ed colleagues who attended the past editions of the Higher Ed Analytics Conference said about their experience.