Tuesday, September 8, 2009

What I would like to see from email vendors...

OK folks, it's dream time...time to lay out the wishes and see what happens...

The email industry is at in interesting crossroads. Current tools are - to be kind - rudimentary in their overall marketing utility to an organization. They tend to be built to keep the geeks (rather than the leaders) happy. While this isn't a bad thing, it's a HUGE limiting factor in the acceptance of email as an important channel for many organizations.

Email systems (both internal and external) are all trying to be good at one thing - getting out the mail. With advances in skill and technology, much of the worry about cranking through a list and sending it out is greatly reduced. With dedicated appliances like Netezza, the issue of query time is largely reduced to rubble. It's list processing, people.

So what can the email community do to increase the value of our precious tool? Here's a few ideas...

(1) Component reporting - There are many pieces to an email, including day, date, subject line and time (among others). Each of these "components" should be readily available via the reporting tool, so that more sophisticated analysis can be performed. You ever try and do a time of day analysis? It's really, really hard - especially if you mail across multiple time zones.

(2) "What's happening now" dashboards - How many emails did we send today? How many people clicked on an email today? in the last hour? What's the status on all of your email jobs (publishing, stuck in queue, etc)? Of the campaigns that are out there generating clicks, how many are from campaigns launched this week versus previous weeks? It might seem simple, but these are the sorts of metrics that make it look like you're in control of your program.

(3) Database metrics - How big is my database (yes, it matters)? Do different regions have different growth patters? Is my database growing or shrinking?

(4) Trending - One of the great weaknesses of email systems is that they ignore the impact of time. For example, how big is my database this month versus six months ago? A year ago? What is the decay curve of an email? How long is the message good for?

(5) Integrated MVT (with tracking) - Everyone talks about multi-variate testing, but it's pretty impossible to do with current email systems, generally because true MVT needs the ability to track an email that is uniquely generated at the blaster (after the campaign is launched). Otherwise, I have to cut up my campaign into a bunch of ugly test buckets...which is not pretty.

(6) Segmentation priority - Let's say I have a campaign that pulls 10 behaviors/products into a segmentation array. Of those 10, which has the most impact? What is the marginal impact of each of the additional variables? Most of our segmentation efforts suffer from the fact that we don't know the impact of adding more variables.

(7) Optimization - What's the right size for an email? What's the right number of emails to send to a person before they get exhausted? We all fight the batch and blast mentality, but our current systems do little to help us battle product marketers who are foaming at the mouth for "a bigger list".

(8) Integration with search and the shopping cart - Epsilon published a great statistic that 1/3 of your email recipients will visit your site directly rather than click on your email. That's a huge impact. We also know that when you launch an email campaign, search activity goes up. Some sort of measurement for both of these varaibles needs to be developed to unlock the true potential of email.

Wow...that's a long dream!

I realize that some of these things may be available. I also realize that it's hard to develop this stuff. But I also know that the people who develop these tools into an integrated package will be the ones who win in the end. There's a lot of email companies out there right now. There won't be as many in 5 years. Organizations have woken up to the fact that email works - the companies who deliver on the above expectations (as well as many more that I haven't listed here) will be the ones who win.


  1. Great list! We also need some good dynamic module reporting, which seems to fit with your call for component reporting and multivariate test reporting. And yes, some have this but it's the integrated package with easily accessible insights that will make an ESP stand out.

    Speaking of which - what are your thoughts on integration with marketing data warehouses?

  2. Integration with MDW's...nice topic for a blog post!

  3. Full transparancy of inbox placement rates. Right now the tools report on "delivered" (either a non-bounce or accepted at gateway) and skip ahead to opens. Missing in there is a big one: Inbox placement.


  4. This is a good list. I'd add the ability to automatically target an ad or message to each user (disclaimer: my company AdRocket does that). This would solve your segmentation array issue and drive up ROI. For example, automatically put the female oriented message/graphic/ad in front of (f) and the male one in front of (m), and of course much more granularity than that.

  5. As a database guy not involved directly in DM systems (anymore) but still deep in the BI trenches, I love hearing the first four points. These are things I wouldn't think of capturing but are remarkably simple to do once it's considered.
    With (6),(7), & (8), I seem to recall that these were the selling points I heard for big, crazy-expensive MDWs. However, with the significant advances in db tech and the reductions in hardware costs, I could see some up-and-coming vendors (those that don't already have an interest in selling MDWs) bringing these features to the table as reasonably-priced add-ons and really start stealing biz.