You slave over your community. The alerts never stop chiming on your phone, your email is the first thing you check every morning, and you know your all-star members better than you know your best friend.
You are a master of online engagement.
So when your quarterly meeting with the executive team rolls around, you’re ready. You hit your presentation out of the park… until the CFO asks how online community engagement is impacting sales goals. Do all those forum posts translate to more revenue? Can you show them those numbers?
Talk about a challenge.
For most of us in the online community industry, we live and breathe engagement. We know to our core that it’s valuable and it does impact business results. But demonstrating that to management with rock solid analytics is tricky.
How do you prove that the countless hours spent on engagement strategies, increasing discussions, and getting more members to log in, had a larger impact on your business?
You show your executive team a regression analysis.
A regression analysis is a statistical model illustrating the relationship between two variables. The relationship is displayed in a scatter graph with a line of best fit, as shown on the right. In the case of online communities, regression analyses can show how engagement activities impact business goals.
A simple regression analysis will show you one of three relationships:
Positive – In positive relationships, the more engagement, the better your business results.
Negative – In negative relationships, the more engagement there is in your online community, the worse your business results.
Neutral – If your regression analysis shows no relationship between variables, then the amount of engagement in your community may not be significantly influencing your business goals.
In most cases, we expect online community regression analyses to have positive results, like the one shown above. That’s because research by major organizations like Gallup and Bain & Company has shown that engagement helps organization sell more as well as retain customers and members.
That doesn’t mean all your analyses will be positive, however. You may find that the more your community members post in your “Help” forum, the less they buy from you. These members are likely struggling with your product, so it makes sense that they would have a less productive relationship with your organization.
Like any analytical tool, regression analyses provide information. It’s up to you to use that information strategically. Here are three top ways community professionals can put their regression analyses to work.
At its core, a regression analysis is a relationship tool. It will show you how different levels of engagement match with different business goals, making it easy for you to spot patterns. You can choose any engagement metric you like, such as discussion forum posts, or you can do an overall engagement score, or level, for your members and see how it relates to your business priorities. In many cases, the higher your members’ engagement, the better your business results. That’s something the C-suite wants to see.
You can do multiple regression analyses to find out how impactful different engagement activities are on your business or association. Once you’ve found the activities with the greatest impact, you can promote those activities in future member engagement strategies. That way, you’ll improve your community’s business results at the same time you’re increasing engagement.
As mentioned earlier, not all regression analyses will show you positive relationships. When you discover engagement activities that are negatively correlated with business results, don’t ignore them. Use them to identify community members who are struggling and then reach out to help. If you can turn their experiences around, you could help your organization increase retention and satisfaction rates.
You can easily do a regression analysis in Excel using data from your online community software and customer relationship management software (CRM) or association management software (AMS.) We’ll walk you through the steps you need to follow, using these two variables as examples:
Engagement Data: # of forum posts
Business Result: total dollars spent
The regression analyses we do with these factors will show us how the number of forum posts published by community members impacts how much they spend with your organization.
Input your data into excel in two-column format. Engagement data should be your first column, business data will be your second column. We’ve included an example data set here, with an extra column for member name. You do not need your members’ names to run the analysis.
In our example, discussion posts are on the horizontal axis and total dollars spent is on the vertical axis.
Excel will do this for you, too. Click on the “+” sign, or “Chart Elements” section of your scatter graph. Check the box for “Trend Line”.
Your regression analysis will show you how impactful member and customer engagement activities are. Use those results to determine what activities “move the needle” most for your organization. Once you have your results, you can take them to your association board or CFO to help explain the impact your online community is having on business goals.
You can also use your results to be more effective as a community manager. Focus your time and member engagement strategies on encouraging the most impactful activities. Or, if you notice negative relationships between activities and business results, keep track of members who perform those activities. Reach out to see if they need help and try to turn the relationship around.
Expert Tip: Regression analyses can be great predictors of future community activity and business results. If you increase the number of forum posts per member from two to five, then you can predict that total dollars spent per member will rise as well. This may give you a better estimate of what your online community can realistically achieve for your business.
Since the simple regression analyses we’ve discussed only test two variables at a time, you’ll get the most useful results by being as detailed as possible. You may find no relationship between discussion posts and dollars spent, but a positive relationship between posts in certain discussion forums and dollars spent.
With that in mind, don’t just test general engagement activities. Test posts in your “Help” forum as well as those in your “New Product Ideas” forum. The “Help” forum would be more likely to show a negative relationship, while the “New Product Ideas” forum is more likely to be positive. The more granular your analysis is, the more likely you are to find actionable relationships between engagement and business goals.
If, however, you want a big picture of how total community engagement impacts a business goal, you can combine activities into an engagement score. Total engagement scores would then replace metrics for individual activities to show how overall engagement impacts your business.
Regression analyses are estimates. They show you the general relationship between two variables, but they’re not always accurate.
All regression analyses have “residuals” or differences between the predicted numbers and what shows up in real life. These residuals are caused by the myriad of factors your analyses don’t consider. For instance, engagement isn’t the only thing impacting how much your community members spend. Their income and personal expenses are also significant factors.
Use regression analyses as estimates, or ways to demonstrate the business value engagement. Don’t view them as hard-and-fast rules.
There’s a difference between correlation and causation. Correlation describes an association between two variables, such as when engagement and sales increase at the same time. Causation means that one thing lead to the other. If we were talking about causation, you’d say you achieved higher sales because of increased engagement.
Causation is very difficult to prove, especially in business and market situations because there are so many factors (those pesky residuals) that you can’t account for. That’s one reason why your regression analyses don’t show causation. They show correlation.
The positive correlation between engagement and business results is well-documented and backed by solid research from reputable organizations. It’s very likely that engagement does directly lead to better business results (causation) but since you haven’t proven that, be realistic in what you present to your executive team.
Regression analyses are incredibly useful. They’re clear, visual ways to illustrate the relationship your online community has on business goals. They’re also great tools for informing your engagement strategies.
Unfortunately, performing regression analyses is also labor intensive, especially if you follow best practices and get as detailed as possible. To get the most out of your analyses with the least amount of effort, perform them at the most strategic times. That may be once every quarter, before you develop your new member engagement strategies, or just before your executive meeting.
And don’t forget that statistics won’t always do your online community justice. The connections your members are making with their peers can be indescribably valuable, helping members get new jobs or feel as if they belong to something bigger than themselves. That’s harder to quantify and demonstrate, but no less worthy of consideration.
Blend metrics and regression analyses with your online community’s other benefits to give your organization a complete picture of how it’s contributing. Such a comprehensive view is what people outside of the community industry need to understand its value.