Intelligent Marketing with Web 2.0

by Luke Donnelly

In the 21st Century, companies will rely more heavily on the artificial intelligence of the Social Web (Web 2.0) to market products and services to people more effectively. Effective marketing relies on intelligent spending, using those precious marketing dollars to reach the right demographic targets with your message. Why pay to broadcast your advertising message to a wide audience, when you can make useful "recommendations" to only those that are likely to be interested, at a much lower cost? This is why companies are starting to provide goods and services using the tools of Web 2.0.

Even today, in the infancy of the Web 2., take companies like London's Last FM. This online music service "scrobbles", or collects data about what you listen to on iTunes, and displays it to other Last FM listeners, and in real-time on your Skype profile. It also collects data about what you've been listening to on the go, by eavesdropping every time you connect your iPod or iPhone to your computer. The service then slowly, over time, builds an accurate profile of the kind of music you like. Even if your tastes are varied.

Using tagging and intelligent software, Last.fm uses your profile to make recommendations about songs or artists you might like to pay to download and concerts you might like to attend, based on your location and tastes. It identifies other listeners with similar taste to yours, and makes recommendations such as "We know you haven't heard any music yet by X, but 85% of people with tastes similar to yours often listen to the new album by X". You can play a sample song by X and watch a video. One click later, you've bought the album, not from a traditional ad, but from a recommendation -- a helpful, accurate recommendation.

"Recommendation marketing" is useful, but it is crucial that the recommendations are accurate. Look at Amazon. When I buy a birthday present for my sister on Amazon, I factor in that she has very different tastes from me. So it's exhausting when I then receive inaccurate purchasing recommendations a few months later from Amazon: "Hey, we think you might like to buy the new book by (insert name your least favourite Chick Lit author here)".

Doesn't Amazon's recommendation service remember that I didn't choose to send that previous gift item to myself? I chose to send it to someone else! So the later recommendation that Amazon makes to me should be based on MY tastes, not on my judgments about gifts for others. There's room for improvement.

Companies know that the social web is big business. A recent business case competition on tagging for taste and suggestion was held by Netflix, with a one million dollar prize! Using the Netflix online DVD ordering service, customers build up a preference list of movies they would like to see. The problem is that customers are not walking around a rental store. The only real estate they have is the computer screen. So if they don't see titles they might like, they don't rent, or they rent less frequently.

Making accurate recommendations about what a consumer might like is a important when managing a business. Why pay for warehouses full of DVDs that no-one wants to see? How many copies of a new European arthouse movie should the company order for their St Louis warehouse, for example, based on the film-watching tastes of their customers in St Louis?

Netflix has their own recommendation engine, but they suspected it could be improved. After years of trying to improve it themselves, they thought there was still room for improvement. So they exposed much of their customer statistics and ran a huge business case competition over three years, that ended last week. The prize of a million dollars went to a team that designed a recommendation engine that could beat Netflix's existing engine by 10%.

But this is just the beginning. A few short years from now, with (and in some cases without) our permission, smart companies will be using Web 2.0 marketing technologies to know exactly what we listen to, what we choose to wear, what we watch, where we like to eat out. But also, what we have not yet watched. What we have not yet listened to. Where we have not yet eaten, but, based on accurate judgment about our tastes, where we would like to eat.

At that point, the common questionnaire will be dead. Web 2.0 will be almost as intelligent as humans, and a huge percentage of company recommendations will lead to purchases. Enriching your experience, increasing companies' earnings, and every marketing dollar spent will yield returns.