The UK's Liveliest Small Business Newsletter
Old archive | New site
|
|
Brick No112: How to convert trials to sales
By Matt Weston, Tuesday 16 November 2004 Protection racket If you didn't receive Friday's brick it was probably because your spam filter took offence at the f-word I kept using. If spam came through my letterbox rather than into my inbox, I'd have to clear it from my front door every 10 minutes. A few months ago, when I returned from my first break since starting this business, that's how it felt. A welcome home party of over a thousand credit card offers, sexually explicit ads, (ahem) pharmaceutical plugs and phishes. I hired protection. My spam cop (SpamSieve 2.2) picks up scores of spam emails a day. It uses a Bayesian filtering technique to identify spam. In plain English this means that it learns from experience. You feed it good and bad messages, and tell it when it has made a mistake. It doesn't automatically delete or block suspected spam, it filters it to a separate folder. If the program wrongly flags a good message as spam (or vice-versa) you tell it, and it learns for next time. And because it learns, the system keeps getting better and better over time. I'll stop right there. This isn't just a plug for SpamSieve. For starters, it's Mac OS X only, and I know only a small percentage of you use that. And neither lesson today is a technology one. How to convert trials to sales Lesson One: The way SpamSieve (and many of the other anti-spam tools on Spamotomy.com) is sold is a lesson par excellence in how to use the f-word. You can trial it for 30 days, or 20 uses if that comes sooner. But as I advised last Friday, that trial isn't pushed as a headline offer. I was sold on the benefits of the program before I even noticed the fr*e trial. The trial simply tipped me to act immediately. And in 20 uses - as you train the system yourself - you find you can't live without it. The more the system involves you, the better it gets at thwarting new spam attacks. I bet the publisher's trial-to-sales conversion rate is sky high. Second lesson from the spam cop Lesson Two: You may have to think laterally here. The alternative to a Bayesian filter a la SpamSieve is a static spam blocker. Instead of teaching your spam filter step-by-step by its mistakes, static spam blockers obey a set of pre-determined rules. Every so often, you can download an update. The problem with this is that as time goes on, the service deteriorates, and the rulebook becomes outdated . . . until the next update. And generally, this is how big business works. When dealing with customers, employees, products and processes, big companies have to adopt a rule book to instil uniformity. They are like static spam blockers: service deteriorates, and learning is lost . . . until the next edition of the rulebook. Contrast this with SpamSieve. "We" are now up to a 99.4% correct identification rate (from an initial 85.4%). Every day that rate climbs. As small business owners, we can be Bayesian. Every scrap of feedback and learning rewrites the rulebook immediately. It's the "Ready. FIRE! Aim." approach, repeated ad infinitum. (See Brick No62.) Remember to sign up: back to top |
|
|
Link to us | Privacy Policy | Disclaimer | ©2005 Business Bricks Ltd
Designed by Mint Digital