The rise of the machines, and the fear associated with it among humankind has increasingly crept into popular culture, in some subtle and not-so-subtle ways. Whether smart machinery or artificial intelligence (AI) is a good thing or a bad thing, especially as portrayed in film and fiction, often depends more upon the intent of the creator and the law of unforeseen consequences than the nature of the intelligence itself.
On a daily basis, the average retailer gathers enough transactional and personal data to feed a growing intelligence network that could be smart enough to function on its own in less than a year of constant-cycle learning. That’s a tremendous amount of data! It takes humans 15 years to amass that level of cognition and ability, on average. Sometimes this “knowledge” is used for good, and because there are limits placed on its use, either by the technology itself or by the circumstances of its use, all is well. The nightmare starts when those strictures and parameters are eliminated, and the machine can “learn” from all sources continually and can act and react accordingly.
The most widely used and easily recognized execution of this is the modern shopping algorithm. An algorithm is simply a comparative database that allows information to have tags attached to it, and when there are several tags in common between two items, one is deemed “related” to the other. This is a simple but very powerful idea. Humans are designed and hardwired to seek out patterns, both visually and in context, but computers are much better suited for this task as they have perfect recall, and aren’t influenced by loss of memory or emotion.
Smart machines that use an algorithm can appear very “smart” to their human users. Amazon.com was one of the first, and most famous, users of a comparative algorithm, when applied to book titles on their bookseller website. Customers would make purchases, the computer would keep track of these purchases, and build a data profile from the tags attached to each item purchased. The tags would then be used to compare these purchases to other books also available on the site, and “recommendations” would be made by the algorithm, based on the number of tags in common. Pretty slick, and with the right verbiage attached to the recommendations, it looks almost like there is a human making the picks and the recommendations.
TIVO television DVRs use this technology to make decisions about what TV programs you like and make recommendations and create recording timers accordingly. Not an infallible system, but it can be remarkably accurate, and it gets better the more decisions it makes and the more data it amasses. Now, extend that capability to association or non-profit membership groups, and as a marketer, think about your annual conference, seminar, or continuing education program. Where does that extension take you?
Why not use an algorithm to help attendees pick conference sessions? Will it improve member engagement? Will it increase overall enrollment? Will it help balance out room set up and class sizes? Can it be used to build tracks or new program offerings in the future that are successful? My research tells me that this type of personalized approach would be well-received by the vast majority of individual association members in a wide variety of industries. The commercial marketplace has gotten them comfortable with the technology, and they understand that the ”Machine” isn’t making life-or-death decisions, merely suggestions based on history, commonality and goals already stated. Generally if it saves attendees time, allows them to navigate a wide spread of data quickly, make some choices effectively and the results are reasonable, I think most event attendees would welcome such a system with open arms.
This type of innovation offers benefits for the organizer as well. Instantaneous feedback of popularity of each session based on purchases, the ability to add additional sessions in the same vein, or to cancel sessions that don’t attract an audience, means your conference department appears to have a solid handle on the needs of the members and can react to them quickly and effectively, with less waste. In short, the data embedded in the algorithm, and the resulting choices it returns, allows for smarter, faster, more efficient product and program development, with less risk, and greater reward.
One or two membership groups have put the power of the algorithm to use with good result. I hope the industry as a whole embraces this use of technology to improve their educational offerings, and for those organizations with more of an a la carte benefit offering, that this same technology can be applied to member benefits as well, providing a highly personalized experience for each member, quickly easily and intelligently. I say, “Let The Machines Rise!”
