Based on results from 1,200 email campaigns that we run for Jungle.com (and some for Argos) in 2000-2003 I have come to conclusion that ideal frequency for us (B2C IT e-tailer) was once a week.

A good question would be what made me think so. Well, I based my conclusions on the following key indicators tracked on per campaign basis:
1) drop off (churn) ratio
2) opens
3) clicks
4) sales (or profits for cases with high variation in margins)
5) costs

Our objective was to maximize 4) which in turn is closely correlated with 2) and 3) while minimizing 1) and 5).

We found ideal formulae for us by running tests which demonstrated that we can get more response overall by mailing newsletter on weekly basis. The key to solution is to test your userbase and this would most likely mean to choose right supplier (unless you doing it in house) that allows for flexible testing (zero or low fixed per campaign costs).

Costs (5) were taken care of by developing in-house mailing solution (by yours truly) and automating as many systems as possible which allowed easy content generation so that we could afford to run 10-15 targeted emails per week using part time effort.

Personally I think email should go out at least once a month - doing it once a quarter will give you headaches with users not remembering they actually signed up for it. This is not to say you need to email more often than you really have something to say to your recepients!

regards,

Alex

Less than once a month sounds too infrequent to me, this is likely to lower open rates as people forget about you.

The 'ideal' frequency depends on how often content refreshes/is useful to the reader, and the repeat purchase cycle of the product your client sells. Weekly emails make sense for groceries (particularly if they contain short-lived coupons), but maybe not for cars. In addition, your client should segment the database and contact best customers more frequently than others with tailored offers (but it sounds like they are struggling to find relevant content even for the monthly email). Once you've established your basic frequency you of course watch opens, unsubscribes etc to get the right balance, as Alex suggests.

Looking at the (surely unrepresentative) sample in my Inbox over the last few months, most ecommerce sites send me a message roughly every 1-2 weeks. I personally think 1-2 message per month is enough, but make them great offers and really relevant.

- Amazon: generic emails monthly, but I also get tailored offers based my order history so contact averages twice a month. They carefully monitor the total contacts so no customer segment's inbox is flooded
- Argos: 2-3 emails per month
- Aveda: every 2 months
- Boden: weekly
- Boots: roughly weekly, although they went into overdrive before Christmas so I got 6 in November and 4 in December which was a bit much
- Carphone Warehouse: once a month (and entitled 'monthly newsletter')
- Gap.com (US): once a week
- Hawkshead: twice a month, but going up before Christmas to once a week
- Lastminute.com: once a week - and they get away with it because they make a real effort to make it fun to read, it won awards etc - they definitely used the newsletter very early on to build the brand, and invested time and effort into it.
- triyoga: once a month
- M&S: roughly weekly
- Dabs: every 2 weeks
- Screwfix: every 2 weeks
- Selfridges: once a month
- Victoria's Secret: every 2 weeks, plus additional mailings during Christmas and sales

There, now you all know where I shop ...

Hope this helps.

Obi

Based on results for The Winchester Diet, I find that every 9 days brings good results.

> - Ask existing subscribers - why second guess it through Unsubs
> and Opens?

I would not call it "second guess" - more like scientific technique to maximise ROI. Same technique can be applied to test various designs, products, marketing messages in the email.

I am all for giving customers choice on how frequent they want to receive emails, however experience shows that people rarely have time or energy to fill in their preferences.

>So we're both right!?

Indeed we are - it is however difficult to explain that to some people who may think that allowing customers to choose frequency of contact would result in drop in number of people "contactable" in period of time most desireable to that person (ie most frequent). I found that using stats that show increase in drop offs and response helps present a much more persuasive argument.