Vendor Relationship Management workshop

Nobody knows you as well as you do. Or do they? Let's run a test. Do you know what percentage of your food bill went to processed products? Or what type of coupons (store coupons, newspaper coupons, etc.) is most likely to get you to switch brands? I bet someone out there knows.

This kind of data mining is the modern companion to Customer Relations Management, which is the science of understanding customers and trying to get repeat business. CRM can offer many valuable benefits, but ultimately the control lies with the vendor, not the customer.

This bothers long-time marketing maverick and Cluetrain Manifesto coauthor Doc Searls. Several years ago he thought up an alternative that would put the data and the control back in the hands of customers, and called it Vendor Relationship Management. He's been pursuing that dream for two years as a Berkman Center fellow at Harvard, and this week he ran the second workshop hosted by Harvard on the topic.

I dropped in and out for a few hours and picked up some ideas, annotating them (as always) with ideas of my own.

Success requires making both sides happy

Most consumers don't think much about how they're being managed. Asking them to take control of their data would come up against blank stares. And naturally, vendors hold on tightly to the CRM techniques and data--no matter how flawed--that they've spent millions to develop. So to go anywhere, VRM advocates have to offer juicy carrots to both customers and vendors.

One approach is to ignore financial benefits to start off, and look for data that customers are already collecting or would clearly benefit from collecting. Customers might want to keep track of their purchases, or what music they've listened to, or some other personal information. A slick data-sharing system with adequate privacy protections could persuade them to combine data with other customers for the purposes of making recommendations or other benefits.

With customer behavior in place, VRM systems could then offer data to vendors, who would rush to take advantage of this new channel for accessing potential customers.

It would be particularly clever to replace costly vendor operations with activities performed by customers. The classic example of this is the displacement of boutiques by department stores. Instead of paying a staff person to take out clothing or shoes, the stores laid out the wares and let customers make their own choices.

Would VRM be better than CRM for either vendors or customers? Well, consider this: today's CRM places data in silos. If I book a hotel in Expedia, I can rate it there but not in hotels.com. Amazon.com knows the books I bought from them, but not the ones I bought at Powells.com. By contrast, in VRM, each customer would maintain a comprehensive data set on himself or herself. The big question is how to get customers to combine that data and offer it to vendors in a way that preserves accuracy and privacy.

Here's a related problem: data maintained by vendors gets out of date. They don't know when you lost your job--or conversely, when you finished a graduate program and received a 30% raise. Would a vendor give up control over data in exchange for a right to access data that the customer keeps up-to-date? I bet they would. What the two sides need is the trust that each will behave honorably. And that requires incentives for honorable behavior. In particular, the customer must be able to learn about abuses of data and then be able to shut off access to offending vendors.

VRM would supplement, not supplant, normal ways of doing business. Companies could offer different sites with different terms of service and privacy policies. Customers who care about managing their own data and shopping experiences would choose the VRM-driven services. One of Doc's recent blogs optimistically suggests "how VRM helps CRM"--and in fact may lead to wholesale changes in marketing and product design.

It really is all about the data

If knowledge is power, that power goes where the data goes. VRM requires each customer to maintain control over his or her data--but how can we ask hundreds of millions of technically naive people to do this? And how do they make sure this data doesn't get lost or corrupted?

One feasible model at present is cloud storage. Customers could sign up with a service that encrypts their data and stores it in Amazon's S3 or some other provider. But the service they sign up with becomes another choke point, in place of the vendor.

Ultimately, VRM requires individual storage and peer-to-peer data exchange. I brought this issue up recently in a blog on social networking. Experimental systems for peer-to-peer data exchange show some promise; for instance, I discussed Jesse Vincent's Prophet with workshop attendees. But the technical and social challenges are formidable. How do people find each other? How do they form a community that can share data in the first place? How do they weed out bad actors and prevent vendors from setting up fake accounts to game the system?

Of course, ordinary web sites and online communities face these exact problems now. To get visitors, online companies spend enormous sums on billboards and TV ads. And when we check Expedia, hotels.com, or Amazon.com, it's completely possible that the reviews we see are fake. The web should be able to do better, and perhaps the solutions can usher VRM in along the way.

Could VRM mean Victim Relationship Management?

A good Vendor Relationship Management could totally change the consumption of goods and services. Savvy customers could:

• Choose the companies with whom they share their data. They couldn't prevent a company from passing data on to a partner--only laws can do that--but they can use a technology such as OAuth to selectively release categories of data. For instance, they won't have to surrender their whole purchasing history to a bank in order to get a loan. The bank just won't know what's being withheld.
• Create new criteria for signaling needs and judging vendors. If hotels don't list the quality of the chairs as criteria, customers who need good ergonomic support while working in their rooms can create their own RFPs (there are services that allow this sort of thing already) specifying that they need a good office chair in the room.
• Search their own data using whatever sophisticated tools arise in the community, thanks to data stored in standardized, programmer-friendly formats.

That's for savvy customers--but what about all the rest? Flexibility brings complexity, and complexity brings the dark side. Many customers are already sitting ducks for phishing and other social engineering attacks. VRM that moves too fast and gets too far ahead of the users can lead them to make decisions they'll regret later.

VRM will have to backed by strong legal protections. This doesn't necessarily mean regulation. A VRM public organization could define criteria for fair terms of services and privacy policies, and place a trademarked seal of approval on sites that follow them.

Where to start VRM?

It's clear by this point in the article why there are no real VRM systems yet, although a few progressive web sites demonstrate bits and pieces of that approach. Much of the Harvard workshop was devoted to finding industries that might be good candidates for VRM. Likely criteria include:

• A highly fragmented market. If there are only a few vendors, they'll have little interest in trying new things that reduce their control, and will instinctively band together to squash innovation in this area.
• A fast-moving, diverse industry whose products vary across many dimensions. This gives customers something interesting to research and rate. Commodities don't offer enough differentiation.
• A product or service heavily used by computer programmers and other technically focused people who will make up the first adopters.

I don't expect to see actual VRM in the marketplace for a long time. Having been following similar peer-to-peer efforts (most of which have shriveled up) since 2000, I'm not impatient. The questions raised by the Doc and the VRM community are important for all of us. He points to efforts going on by a number of online projects, and says "I think we'll see some good examples within a year or so." But even if VRM remains a thought experiment, it will pay off in spades.