How to Calculate the ROI of Online Communities

By Richard Millington

ROI People

One of the lesser-seen benefits of creating a community is the option to reduce the cost per call by improving agent handling time. This is the ability of customer service agents to handle incoming calls and emails more efficiently, thus reducing the overall cost per call by increasing call volume. This can be achieved by improving agent productivity, i.e. improving the average handling time through superior product or customer knowledge.

Agent productivity is defined here by the average time spent handling time on calls. If, on average, an agent spends 300 seconds (5 minutes) on a call, we can assume they could take around 70 to 80 calls per day.

However, if, through superior knowledge and product information provided by the community, this time could be reduced to 4 minutes, they could take 95 to 105 calls per day.

A community can increase agent handling time by providing better quality information to the agent (e.g. unearthing new issues and finding solutions faster or better than existing material available).

To measure the agent handling time improvement, you need:

  1. Average handling time of agents ‘not exposed’ to the community.
  2. Average handling time of agents ‘exposed’ to the community.
  3. Cost per call via current methods.
  4. Total number of calls received per year.

Community ROI Template

You can add these values into this spreadsheet or follow the process below.

Step One: Measure the Impact

The impact can be measured in one of three ways:

  1. Trends before and after. Measure the average handling time before and after the community has been developed to look for clear impact. This will show possible improvements but will also include any general improvements (or deterioration) in average handling time which might have little to do with the community.
  2. Correlation. A more accurate method is to correlate the amount of time a staff member spends reading the community material with their average call handling time. In practice, this data can be difficult to gather and does not establish causation (e.g. are call agents most likely to resolve problems quicker or simply more likely to visit the community?)
  3. Split test. The best solution is to run a simple split test. Randomly assign two groups of a call center into one of two conditions. The first is the control condition. They continue to perform as they are today. The second group is the community condition. They are given access to the community, summaries of key content, and taught how to use the community. After a defined period of time (e.g. three months), measure the difference in average handling time between the two groups. If the average handling time for the control group remains at 311 seconds and for the community group it drops to 303 seconds, this is a clear 8 second difference, which we can attribute to the community. If the average call lasts 311 seconds previously and now lasts 303 seconds, this is a 2.57% improvement in efficiency.


Note: The third condition would also entail additional upfront training costs which may need to be factored into the equation. This doesn’t include the natural diffusion in knowledge. Once knowledge is acquired, it is naturally likely to diffuse among customer service agents regardless of direct exposure to the community. We’re assuming these two cancel each other out. The reality is probably more complicated.

Step Two: Calculate Cost Saving Per Call

Now we need to determine the total value of this efficiency.

If the average cost per call before was $7.01, we now multiply this by 2.57% to determine a new cost saving per call of $0.18.


This means in year 1 we now save $0.18 on every call received by the customer service team. We can use this to determine the total cost saving via improved handling times.

Step Three: Calculating Total Cost Savings

This is a simple calculation of multiplying the number of calls received by the cost saving per call to measure the cost saving of each. We’ve shown the full template below.



  1. Agent handling time is the amount of time a customer service agent spends on a single call/problem.
  2. A community can improve agent handling time by increasing the amount of information available to the agent (unearthing and solving new problems).
  3. You need to run a controlled test to measure the impact of the community on average handling times. Once you have this data, it’s easy to multiply the % difference by the cost per call and then the total number of calls per year to find the total amount saved by improved agent handling times.



 LinkedIn LinkedIn

 Google Plus Google+