Archive for October, 2007

3 pressures to integrate

Friday, October 19th, 2007

At the IDM members evening “View from the top: new advertising opportunities” last night, we were treated to an extensive and in-depth review of the advertising landscape, courtesy of the IPA’s latest Bellweather report, presented by its Director General Hamish Pringle. There was a lot to think about, but I was struck in particular by reference to the fact that marcoms has become “salami sliced into niches”, making it difficult to manage and derive the maximum benefit. According to Hamish, this creates three pressures to (re)integrate:

  1. Marketers demand for integrated communications – emails linked to microsites, advertising with personalised response, end-to-end lead tracking
  2. Procurement demand for efficiency – let’s face it, we all want efficiency (but we blame Procurement when it gets too hard)
  3. Digital media requiring creativity with technology – not natural bedfellows

What does this have to do with marketing insight? Well, I’m bound to say it of course, but data, analytics and intelligence lie at the heart of this re-integration. Not that there aren’t other elements of course (creativity, technology), but the operational underpinning is crucial to making it happen – more gritty stuff of marketing perhaps?

Why it’s important to integrate your database and email marketing

Tuesday, October 9th, 2007

We use external email service provider eMarket2 for all of our email marketing execution in EMEA, whose web-based, template-driven self-service SalesTalk system is very easy to use for the creation of standardised HTML emails. Using a provider like this has lots of benefits, not least deliverability, as they take care of black lists, spam honeypot addresses and other considerations. Contrasted with the uncertain deliverability of our email marketing in the US, it’s seems to be working well for us in EMEA.

eM2 would of course very much like us to utilise the database functionality of their system as our primary marketing data repository, but this just doesn’t fit in with our wider requirements and workflow. As such, we need to ensure a good level of integration with our Onyx system to maximise efficiency and the best-in-class data management. Again, eM2 have an application programming interface (API) which they’re keen for us to utilise for this purpose, but the idea of trying to initiate a project of such sophistication in the midst of the prevailing SFA review isn’t compelling! So, I’m putting a slightly more ad hoc process in place involving batch data transfers and overnight updates, all of which should be automated. Why go to this effort in the first place though?

  • Email lists should be refreshed from your marketing database for each piece of activity, ensuring it reflects the most recent data and appropriate selection possible. This is particularly important for observing opt-outs and avoiding emailing old or invalid addresses, which affects your spam ratings.
  • It’s important that all outbound activity is tracked in your marketing database in order to build up the touch history of your marketing communication. Pulling emails lists directly from the database is likely to be the most efficient means of achieving this.
  • Delivery results, including bounces, opens and click-throughs should be returned to the database as part of your tracking. Bounces are crucial for flagging out of date addresses and individuals that have moved on, together with the spam considerations just mentioned. Opens and clicks let you track the behaviour of your recipients, which can be utilised in future activity. For instance, you could differentiate messaging for individuals who previously opened a message or visited a landing page.
  • Depending on your unsubscribe mechanism, it’s also crucial to ensure that opt-out requests are fed back to your database in a timely manner.

With these processes in place, operating largely transparently to our marketing programme managers, I’m confidant that email marketing effectiveness will improve over time.