In my last post I talked about the new buying process for business software and the role of influencers. One of the new influencer types is peer networks, and since that post I’ve gotten quite a few questions on this category and how it operates. I want to spend some time talking about the general concepts which will be colored somewhat by the specifics of the one example that I know the most about, G2 Crowd. If you read the last post (or go back and do it now), some of this will be redundant, but I want to lay this out in some detail to get all the topics out clearly.
Peer networks fall into three distinct categories, public social networks, private social networks and peer review platforms. Public social networks include the obvious ones like Facebook, Twitter, Pinterest, LinkedIn, etc. and also public communities sponsored and/or managed by an organization. Organizations that could have a public community include any business with a customer, product and/or partner community and professional / industry / special interest organizations like Project Management Institute, and American Accounting Association. Most of the company sponsored / owned communities are built around peer to peer product support and customer service, although there are other types that range from loyalty programs to game driven sites for customer / company interaction. For other professional type communities the draw is mostly networking and education, although it could also provide connections and information for developing new business and personal opportunities, and probably several other things I didn’t think of. All of these networks have the common thread of connecting people around common subjects or goals, and often serve as a forum for exchanging ideas and opinions and asking questions. Obviously that’s how the public networks fit into the influence ecosystem, they enable the connections people can and do leverage for asking for information and recommendations on products and vendors, or in other words the online equivalent of offline informal social interaction. Advice from colleagues and peers always ranks high as an influencer in software purchases and this is particularly true when the source is both in the network so already trusted, and meets the “like me” test. In any influence source trust and credibility are the highest criteria but followed closely by context (which is both timeliness and similar experiences).
Private social networks are very similar to public social networks in the way they function and the amount of influence they exert. The one exception is that context is likely to be higher in a private network, simply because of the nature of the network. Private social networks are just “gated” communities that are restricted in who can participate and usually around the purpose of the community. An example of a private social network is a patient – doctor network at a hospital for cancer patients and their families, or a partner network for a company and its partners like value added resellers (VARs). Functionally it is the same as public social networks, a source for highly trusted peer advice.
Peer review platforms come in many forms but share several characteristics; they all are speciality networks, that is, focused on a specific topic or set of related topics; they collect direct feedback from people that have (or claim to have) direct experience with the specific topic of the network, and they should have a high level of transparency to engender trust and credibility in the content. Beyond that there are many variations to the basic model. For this post we’re focused on business software peer review platforms, like my current employer G2 Crowd, and the ways that the content (reviews and ways to interpret the data) can be used to assist in the purchasing process. The basic process works pretty simply, the community platform establishes an orderly method for collecting and presenting reviews and other information on business software that is categorized and organized in ways to make it easy to find for other people who are interested in the software and/or the business problems the software can help resolve. The site can provide additional information about software vendors, either developed independently or provided in some monitored way by the vendor themselves. In the previous post I listed and explained 6 requirements that online business software B2B peer review sites need to meet, so I’ll just reiterate those in short form:
- Robust online purpose built platform
- Sustainable process for getting reviews
- Methodology and technology for vetting reviewers and collecting detailed reviews
- Underlying structured way to categorize business software products in an industry accepted and recognizable manner, and match reviews to that category structure
- A positive user experience for reviewers and consumers of the reviews
- Value added resources to assist buyers in finding and evaluating business software
There are several challenges in meeting the requirements listed above, beyond the obvious technical and design issues around the platform and the user experience. I will say that having a robust and advanced platform makes most of these other challenges easier at least. Anyway, what are these “challenges”? There is one overarching challenge that must be addressed in everything that is done, that is, building and maintaining trust and credibility among reviewers, software buyers and software vendors (or said another way, in the community). If you can’t do that then the network, and the business behind it, will not be successful. The rest of the challenges generally fall on two sides of the process, some related to getting valid reviews, matched to the correctly categorized products and vendors; and the other set of issues around getting products categorized correctly so that they can be analyzed against products that solve the same business problems, and then provided in some way that helps compare products and vendors to help the buying process.
I could talk about a number of areas related to trust and credibility but to keep this simple here are the key considerations:
- The site and the business must operate in a very transparent way to remove doubts about undue influence from vendors, gaming of reviews, unfair rankings, etc.
- The safeguards on the review process must be very strong to prevent or head off gaming the system, or even the perception that the reviews for some vendors are inflated due to undue influence. These safeguards should include the use of technology to assist in the risk management process as well as other platform tools to make the process bullet proof.
- Make best efforts in all the processes, including categorization of software products and any comparisons of products to “get it right”. Within reason the processes are designed to avoid miscategorization and if it happens it gets corrected quickly.
- Every effort is made to prevent the misperception of any business model conflicts.
- Ensure that any issue that is raised or problems that are discovered are dealt with quickly and transparently.
So on to the problem of getting reviews and keeping those reviews as current as possible. There are several ways to go about “recruiting” reviews but these four basic categories seem to be the most effective:
- Site traffic – organic traffic assisted with search engine placement (SEO) along with gamification and incentives to reward participation in specific and desirable categories. It’s important to note though, that incentives must be for participation only and completely removed from the content of the review.
- Referrals – from community members including vendors (but from vendors the solicited reviews must not be incented in any way nor can the vendor ask for only positive reviews).
- Networking – outreach on public social networks like LinkedIn and in the member community
- Partnering – outreach through a mutually beneficial partnership for trade organizations and other private networks
Recruiting is the first part of the challenge, the second and equally difficult, is reducing the risk of improper reviews as much as possible (I’d call this “best effort”). If there are improperly influenced reviews on the site, or even if there is the perception that some reviews are not clean, trust is broken and the damage is very difficult to repair. Vetting a review includes weeding out reviews that come from vendor employees and family, competitor employees and family, and any improperly incented reviews from anyone. Partner reviews are also a bit tricky, they can be very useful but also tend to be self serving. One approach is to allow the partner reviews but not use them in ranking products and vendors. Valid identity is important as well, even if the site allows the reviewer to remain anonymous on the publically viewable version of the review. Identity is one of the keys to vetting the review and so reviews from people whose identity is in question should be excluded.
Vetting can be done in several ways, and a combination of these methods seems to be the best overall insurance against false reviews. The process requires some sort of human QA, even if technology is used to asses risk. Risk assessment algorithms can be developed that can raise the confidence level of a review considerably. A formal and established QA process that is enforced is needed, with both technology assisted assessment and other forms of research. Ultimately the assessment of usable or not is extremely complex and more art than science, so the final decision is made by a QA expert. Validation can also include the use of social networks, which can be a good check on identity and on potential for actual use of the software (a credibility check).
Categorization of products is also a critical part of the process, with correctly categorized products supporting easier access to reviews and more effective methods for ranking products to assist buyers. There are probably many ways that would work for category design, but no matter the specific method they need to be structured, well defined in a transparent way, matched to industry accepted terminology, and regularly reviewed to keep them current with changing technology trends. The categories generally are hierarchical and work in a tree structure with parent – child relationships to allow for individual products that are often bundled / sold together as a solution or suite. Determining product to category assignment can be difficult at times, particularly with products that are broader than a single category or only sold as a part of a suite, so the process and rules must be clearly and transparently defined. Categories are determined by the function of the product, not how it can be (or might be) used, to prevent confusion or prevent a single product from appearing in multiple categories. The categories must also take into account that buyers will be using them to find products that solve business problems, and therefore can’t be misleading, unclear, or otherwise confusing to the buyer. A big part of ensuring that they categories can be used effectively is publishing and maintaining the category definitions, and having a process to review product categorization if it is challenged by a vendor, vendor competitor, user, buyer, etc. A sanity check can also be applied to whatever ranking / rating system is used, to keep that comparison limited to like products that service like customers.
To assist in keeping the comparisons “like to like”, additional segmentations are also important. Buyers will want to compare products that, besides being functionally similar, are suited to their company size, their growth plans, and their industry and its unique compliance and governance requirements. Many software products are horizontal, that is, can be used in multiple industries, but some are built to only operate for a specific function / need in a specific industry. Those clearly vertical products are relatively easy to categorize together for ease of comparison. Horizontal products can be a little more complicated since they might also have offerings (versions) that include some vertical functionality. For comparison these products generally are compared in the horizontal category, but the vertical variants must be called out to assist in software selection. All products need to be segmented and presented by appropriateness for use in a business size segment (small and medium business, large enterprise or some other classification). Some products might appear in either segmentation but it is still important from a buyer’s perspective to give them a view that is most like them to help them find the best fit for their specific requirements.
Products also need to be segmented into delivery / licensing model, since many buyers will be looking for cloud based solutions, or at least want to evaluate some traditional on premises solutions against the SaaS offerings available. This can lead to a bit of confusion unless those segmentations include clear published definitions for what is and is not considered in each segment. Segments include public cloud (SaaS) applications, platforms (PaaS) and infrastructure (IaaS), private cloud solutions, hosted / on demand solutions and other on premises solutions.
The last part of making review data useful to buyers is to provide some sort of comparison of the products in the same category and segmentation. Since this is such an important part of the usefulness of the site and its data, I’ll save it for a separate post, so I can go into a little more depth (and also because this post is already very long).