Customer communities and discussion forums have emerged over the last two years as a popular new support channel. Adoption of customer discussion forums by SSPA members grew from 36 percent in 2007 to 49 percent in 2008. Unfortunately, there has been too much emphasis by community platform vendors on the trendy aspects of social networking and insufficient messaging on the business value of an implementation.

As a result, companies that are about to kick off a community project, or are in the early stages of an implementation, are being pressured by management to cost justify the project before proceeding. Return on investment (ROI) for communities comes from both hard and soft metrics, but few guidelines to realistic results exist. Companies implementing an online community or discussion forum need to make realistic ROI estimates before continuing, and ensure all the report elements and required integrations are in place to enable accurate ROI analysis when live.

Web 2.0 Buzz Fades in Favor of Concrete ROI
Over the last few years, a number of industry analysts and other experts have published books about how online communities and social networking are reshaping customer expectations and the overall customer experience. Never mind that the intended audience for most of these books is marketing…the buzz about Web 2.0 has influenced every department across the enterprise, including customer support.

With SSPA benchmark data showing self-service success continually trending down,¹ companies eager to boost adoption of unassisted channels quickly gravitated toward discussion forums as a way of deflecting more assisted incidents, as well as fostering a customer community in the hopes of increasing customer satisfaction and loyalty. In the current economic climate, companies are only receiving funding to initiate or complete projects that have a guaranteed ROI, and SSPA members considering a community deployment now are faced with a number of challenges.

• The ROI story is vague. As the ROI for communities has in the past been a supporting message, not the primary message, the best-practice metrics to prove ROI are not well understood, and some vendor-supplied ROI models factor in assumptions that overstate cost impacts.
• Customer adoption is a wild card. As with any self-service project, convincing customers to use the system is a big challenge. With both consumer and enterprise customers being harassed to join so many personal and professional networks, “community fatigue” is setting in, impacting the propensity of customers to join—and be active in—yet another community.
• Project costs are higher than expected. Though freeware for forums and blogs seems ubiquitous, the cost of an enterprise-class community platform remains a major investment, with upfront development charges and ongoing page view subscription pricing. And to properly track ROI metrics, the community platform should be tightly integrated to multiple systems (knowledgebase, self-service, agent desktop), adding additional cost, time, and complexity to the project.

Accurate Assisted Support Deflection Difficult to Calculate
As with any self-service technology, the primary ROI story for customer communities is deflection of assisted support incidents. If customers are able to resolve their problems by posting questions and receiving answers in a forum, or they can find their question already posted and resolved by others, the overall number of assisted support incidents should go down. Vendor ROI models and calculators look at the amount of traffic to the community and the number of threads viewed, and estimate the number of customer problems resolved. Though seemingly straightforward, actual ROI analysis is complex.

• B2B/B2C differences. The publicized stories with big deflection numbers are all from consumer-facing companies, and the results from B2B environments will not be as dramatic. Consumer companies offer few, if any, free assisted support options, and free support that is offered is often characterized by long hold times and lower-skilled agents. For consumers, making use of self-service and any available forum is a necessity. For B2B environments, in which assisted tech support is easily available according to the terms of the maintenance agreement, active communities of system administrators are more likely to trade tips and tricks, not solve each other’s complex technology problems.
• Supplementing existing self-service. Frequently asked questions in your knowledgebase also will appear in the forum. Customers who visit your forum first and find an answer are not necessarily deflected calls; they may have tried the knowledgebase next and found the answer there. Since few companies have introduced a federated search, in which a single search box will find matches in the knowledgebase and/or the forum, cross-promoting these two separate self-service initiatives remains a challenge.
• Lack of integration. Most companies have different systems for incident tracking, knowledgebases, and forums, and little integration exists today between these disparate systems. Unless you are tracking click streams to make sure a customer who clicks “this solved my problem” in a forum doesn’t call or e-mail you an hour later, the accuracy of these numbers is suspect.
• Customer intent. The biggest flaw in the vendor-supplied ROI models for communities is an assumption that any question answered in the forum saved a phone call or an e-mail. People access forums for many reasons: because they have a support issue, because they are curious, because they are bored. Other customers browse forums as a training tool or to resolve a curiosity question they would never have called you about. Assuming every happy forum visitor is a deflected phone call is a gross overstatement of community success.

Creating Realistic Expectations: Start with Channel Volume
The best way to prove that a discussion forum is deflecting live agent interactions is to monitor the volume level of every channel to detect fluctuations when the new channel (a discussion forum) is introduced. What was the impact on phone and e-mail volumes, as well as self-service usage, after the launch of the community? If community traffic accelerates, but there is no decrease to other channels, customers are likely satisfying curiosity questions with the forum, not technical support issues.

Luckily there are documented examples showing that forum adoption has decreased volume for other channels. One example, from Cisco’s Linksys division, was highlighted last year in SSPA research.² When Linksys decided to eliminate the e-mail channel for customer support, 80 percent or more of e-mail volume transitioned to an unassisted channel: the discussion forum. There was a 10 percent increase in Web chat traffic, but zero impact on phone volumes. By showing a shift in traffic, Linksys was able to clearly document the ROI impact of their Lithium community implementation.

Another community platform vendor, Helpstream, offers a full customer service suite including incident tracking, knowledge tools, and Web self-service, in addition to discussion forums. Having all of these elements pre-integrated makes it easier to track volume fluctuations across channels. As a software-as-a-service (SaaS) vendor, Helpstream can track the overall distribution of incidents by channel for all customer implementations. The current distribution is:

• 47 percent of customer issues are resolved via tech support engineers.
• 37 percent of customer issues are resolved using the self-service knowledgebase.
• 17 percent of customer issues are resolved using the community discussion boards.

According to the current SSPA Benchmark, 19 percent of incidents, on average, are deflected to Web self-service, meaning 81 percent of incidents are handled via assisted service. Your current volume by channel should be well understood before launching any new self-service initiative, including a discussion forum. With your baseline established, increases/decreases by channel can be documented as the new forum goes live. So what are realistic expectations for assisted support incident deflection? Based on reviewing a number of case studies:

• A successful discussion forum, once fully deployed, should resolve 10 percent of overall support volume.
• Companies with highly engaged and collaborative customers will find higher deflection, with an estimate of 15 percent. For B2C environments, this number can easily reach 20 percent.
• For the first year of an implementation, particularly for B2B support, a five percent deflection is a realistic goal.

With average Level 1 incident costs ranging from $20 to $45, even a five percent deflection can create a large payback, possibly cost justifying the project in the first year.

The SSPA Recommends
The majority of early success stories for online customer communities came from the consumer world, primarily communications, media, and consumer devices. As this article has illustrated, enterprise support groups have a number of unique challenges that make the ROI for communities less reliable than for consumer firms. However, B2B success stories are emerging, including a new SSPA Star Award category for Best Online Support, and SSPA Research will continue to report on findings moving forward.

A key point to keep in mind is that, like other areas of service and support, success is about process, not technology. Regardless of the bells and whistles of the platform you choose, you will not be successful without the right processes in place. Pay particular attention to your plans in these areas:

• Promote, promote, promote. We’re all being stretched even thinner than usual by the recession, leaving less free time for online community participation. To recruit members and to keep them coming back, ongoing promotions are required. Be sure to feature your discussion forums, along with all self-service options, in each customer communication, including phone calls, chat and e-mails, media advertising, press releases, and bills.
• Survey, survey, survey. Until better industry baselines are established, especially for B2B companies, surveying is critical to understanding forum adoption and success. Why did they visit the forum? What did they learn? Did this really avoid an assisted interaction? How does it supplement self-service? Will they be back, and will they recruit others? It is critical that you understand the difference between demand for self-service and potential assisted service incidents—not all visitors to self-service or the forum have an issue for which they would request assisted support. If you erroneously assume all self-service traffic is potential assisted support traffic, you will overstate the ROI potential for a community.
• Integrate, integrate, integrate. Having a process in place to migrate strong content from a discussion forum to a knowledgebase is key, and having integrations between the community, knowledgebase, self-service system, search, and incident tracking system all streamline use and better enable accurate reporting on success. Look for community vendors with more modules (knowledgebase, search, incident management) than just a forum, or select a provider with an existing partnership (and hopefully a pre-built connector) to the tools you have in place.

Endnotes

1. For more information on self-service success trends, see the December 2007 SSPA Accelerator, Web Self-Service 2008 Trends.

2. For the full Linksys case study, see the June 2008 Best Practices Report, The Challenges of Tech Support via E-mail: Linksys Ends E-mail Support, Successfully Migrating Traffic to Forums.

About the Author…

John Ragsdale is vice president of technology research for theSSPA. John spent 10 years managing tech support operations before moving to Silicon Valley where he held product management and marketing positions at e-service and CRM vendors. He spent five years at Forrester Research as vice president and research director before joining the SSPA. John may be reached at jragsdale@thesspa.com.