A knowledge community needs to consist of diverse people with independent perspectives. Naturally, the most valuable perspectives to understand are those of future customers. But who are the future customers? Do they all think alike?

Diversity must be pursued along several dimensions:

Community Role – Maven, Creator, Proxy, Conduit, User
Business Role – Customer, Vendor, Expert, Advocate
Disposition – Enthusiast, Supporter, Acceptor, Skeptic, Antagonist, Disinterested
Functional Role – Technical, Network, Economic, Strategic
Role in CommunityBuilding – Coordinator, Sanctioner, Connector, Salesperson

Community Building Principles

The objective of Community R&D is to understand how customers think, what their unmet needs and desires are, and why a person would adopt a new offering.

Not all individuals are equally important. To span the knowledge network, it is best to harness natural social networks. People know other people. Not only do they know them, they know of them – what they’re good at, what they know, etc.

To explore this network, the priority is on diversity of perspective and depth of dialog. All customers are not created equal and it is imperative to understand their key differences. Not the difference in behavior, but of motivation.

This is not a statistical approach. It is an approach to answer the question, “How and why do people think, choose and form opinions?” It does not answer the question of, “What percentage of people like the product I’m creating”.

This first question can’t be answered using surveys, focus groups, or even ‘Voice of the Customer’ techniques. Knowing the answer to the first question must precede any effort to understand the second – finding the relative proportion of customer types in the future market place.

The typical number of participants (or ‘community members’) required lies between 20 and 50, depending on the breadth of the area being explored. There is no fixed number because of three attributes of this approach:

1. You know who you are talking to - A structured process is used to identify participants, rather than a random sampling.
2. The questions can’t be known in advance - In-depth dialogs get at underlying, tacit knowledge and behaviors not possible through surveys, focus groups or large volume techniques.
3. The process uses continuous feedback - Process metrics allow seeing the point of ‘diminishing returns’. When additional engagements result in little or no new knowledge, you can safely terminate the process. Experience has shown this happens in the range of 20 to 50 core community members.

Building the Community

Community members are identified in two ways. The first is through active outreach (i.e., recruitment, research, reputation, etc.) and the second by reference from other community members. A key part of the engagement process is determining the characteristics that define a person’s role in the community. There are a set of defined categories (based on Inovo’s research and that of others’) that determines where a participant fits. These categories are along the dimensions of the person’s role in the social network (connector, maven, etc.), their attitude (skeptic, supporter, enthusiast, etc.), and influence (thought leader, promoter, antagonist, etc.). It is usually possible to profile a person and identify others needed to round out the aggregate perspective.

The community is typically comprised of:

• People in the company who can speak for the customer
• People in the industry ecosystem – individuals from vendors, partners, etc.
• People in complementary industries including competitors and alternatives
  
• Customers – a diverse variety of people who will directly experience the product
  

For a product that is ‘new to the world’, or one that could open a market that lacks historical data, the optimal method of performing market research is through outreach to a ‘Community’ – the precursor of a market. The Community includes those whose needs and desires shape our view of what is valuable in a solution, and how to influence its adoption as a product in a new market.

Directed vs. Random Sampling

The determinist side of Community Building comes from its structured targeting approach. Community building hinges on selective engagements to out-perform larger, randomized consumer sampling methods. By exploiting social network theory,[i] it is possible to span a broader space with far fewer samples than a random approach. This is possible because the objectives are discovery and learning, not statistical validation.

Consumer experiences are shaped by social context as well as by exposure to a product or service. To uncover needs and desires, it is best to span the social network. The network is the conduit through which consumers learn from each other, as well has how researchers learn about consumers. When the network and the social context are well understood, then statistical sampling is an adequate research tool. When the network and context themselves are unknown, then coverage (both direct and indirect) is most important.

The network shown below in Figure 2 is a diagram showing a real social network with lines connecting direct interactions. The ten nodes highlighted in red are drawn from a random sampling. Their indirect contacts (their immediate neighbors) are shown in green.

To be more efficient, however, we can exploit certain structural constants that social networks exhibit. These include:

• Super nodes – people with direct exposure to many others
• Weak links – people who bridge sub-communities

By focusing the engagement process on these community members, it’s possible to dramatically increase the secondary coverage. Figure 3 below shows intentional targets in red, and immediate neighbors in green. This example shows the efficiency possible if the super nodes and weak links are identified and leveraged.

Simulations[ii] have shown that exploiting the effects of super nodes and weak links leads to dramatic reductions in the ‘degrees of separation’[iii] in social networks. This is exactly what is required to span a network – even in a cursory manner – for rapid learning. The qualitative claims of network coverage have been substantiated empirically through real-world applications of the methods. The efficiency of the process is illustrated by comparing Community Building to results from traditional, sample-based research. In actual projects, we have seen directed sampling methods take 1/3 to 1/5 the effort, with more insightful results, when compared to statistical methods.

Market adoption is a complex process. While it isn’t possible to predict adoption rates of new-to-world technologies, it is possible to understand who cares, why and by how much. The output of Community Building includes Community Maps, Community Member Profiles, Needs & Desires Maps, and Persona Models. Combined they give the knowledge to think intelligently about the future and make well supported commitments for new product development.