The dangers of clever (email) marketing?

Time for me to be the grumpy sod at the back again. (There's always one).

There's a lot of talk about super new targeting techniques based on subscriber behavior. Here's an example: a customer places a product in his shopping cart but fails to complete the transaction. This triggers an email to him which offers 5% off that very product if he returns to complete the purchase.

In other words, sophisticated email marketing tools track behavior online and send out customized emails based on this behavior.

It's great stuff. The more targeted and relevant the email, the more likely you get a response. This kind of on-the-fly customized email program has a big future ahead of it.

But I never read much about the potential issues associated with such approaches. I write this as a recipient of triggered email, not as someone who has done this kind of email marketing (so feel free to jump in if you have and disagree.)

1. Technical complexity

It's very easy to describe the concept of behavior-driven emails, but nobody says much about implementing it. Your web analytics needs to talk with your email system. That sounds like some heavy-duty technology or upgrades. Or are there integrated off-the-shelf systems available that can automate all this for you?

2. Privacy and permission

People care about privacy and preserving their anonymity, as you can tell from the uproar that accompanies any public discussion of cookies, spyware and similar. Is there an issue here? Will some people object to their favorite shopping site so obviously tracking their activities? Or don't they care?

Plus: permission to send these kinds of emails needs to be built into your opt-in process.

3. Conditioning

We already know that discounting practices can condition customers to expect never to pay the full price. Does conditioning matter here? If abandoned shopping carts always lead to an email coupon, will we train shoppers simply to load up their shopping cart, leave the site and wait?

4. Rule-setting complexity

Then there's the time and effort required to set up all the rules that determine when and what emails are triggered. Some examples:

Timing: how long after the trigger behavior do you wait before sending the email? How long do you wait before you assume that shopping cart really is abandoned?

Frequency: how many times do you allow a trigger email to go out. If I browse, but don't buy, get an email, browse again, but still don't buy, what happens? Another email? How many times or for how long do I have to browse before it's worth intervening again?

Priorities: suppose I look at 25 different products, but don't buy. Which of those products gets the focus in the next email you send me?

False interpretations: since the whole process is automated, how many ifs and buts do you have to give the system? What degree of error can you accept in terms of the system interpreting behaviors poorly.

Here are two real-world examples from an online book retailer that likes to send me emails with recommendations based on previous purchases. (Which is great in theory.)

I once bought a book which helps your kid to learn the alphabet. Two years later they tell me since I bought a book about teaching your kids the alphabet, I might like to know there's a new book out on the same subject.

Whoops: are they suggesting my kid needs more than two years to learn the alphabet?

I once bought a book on Buddhism. Some people who bought books by that author apparently also bought books by another author. The recommendation for that second author? "An Introduction to the New Testament for Catholics."

Whoops: I'm not Catholic, not even close (hint: check the first book topic).

I use these examples simply to illustrate how automation can bring about errors. Of course the net effect of their strategy (mostly the system gets it right) is positive. But is there any way to avoid these errors?

I write this not to appear negative about the potential of triggered emails, but because I'm fascinated at how these issues are tackled by practitioners. If you have any opinions, let me know and I'll publish the polite ones.

Sign-up for the Email Marketing Reports NEWSLETTER
Twice a month, free, packed with email marketing advice and all the posts from this blog.

Email:      First Name:     
    More info and sample

 via feed | via email | LinkedIn | Facebook

The main website for everything email marketing
Email marketing search engine
Introduction and email advertising basics
Service listings and sources of statistics
Advice on email design and delivery issues

Top 21 email marketing info sources
Outlook 2007 and HTML design
42 HTML email design resources