I’ve just received a personal invitation to take part in a specially selected offer that’s been chosen just for me and me alone! I know all that because it says so on the email that’s come into my inbox this morning. It’s so very kind of the retailer in question to bear me in mind. I do appreciate personal recommendations for goods and services that I might otherwise have overlooked and missed out on.
Now, the only question is: as a 6-foot tall man, am I really going to fit into the rather lovely lace sun dress that’s been recommended to me?
You know what? Maybe, just maybe, it’s not that personal a recommendation after all? That seems the only likely explanation, short of the retailer making some rather bold assumptions about me based on my shopping history!
As it happens, I know that Stuart looks exceptionally fetching in one of the Scottish tartans he is entitled to wear but a lacy sun frock? The mind boggles at the prospect. (Personal note: I too am entitled to wear a tartan. Don't go there!) But on a serious note, bad and sometimes wholly inappropriate recommendations are only the tip of the marketing cock up iceberg.
Eric Wittlake points out no less than SIX areas of error where marketing technology is not living up to the promise:
- Dynamic copy that isn't dynamic - assembly of random copy
- Too many segments, not enough content - insufficient content to populate the framework outline
- How much does half a banner cost? - autoplaced banner that wasn't resized
- Nope, that's not me - broken lead scoring redirecting names to the wrong user
- Dear , - a common one where the name field isn't checked for content (or accuracy)
- The Autoresponder - via Twitter, Bank of America’s recent string of automated twitter responses. If you haven’t seen it yet, Eksith captured the original set of exchanges.
No, the answer isn’t to run away from automation, but you will need a new approach to QA, one that doesn’t hinge on reviewing every output. As these examples show, almost all from companies that should know better, it is far easier for a problem to slip through when the technology is now creating the marketing your audience sees.
We don't know the incidence of these failures but the fact one digital marketing expert had no difficulty in assembling these examples speaks volumes to the state of the technology's true ability to deliver what individuals want. To quote Lauchlan:
With that in mind, we were delighted when a recent Lyris/Economist Intelligence Unit (EIU) survey validated our thinking on this subject and supported a personal thesis that the aching for the bleeding edge modernity presented by so many in marketing is overcoming a much needed pragmatism of approach.
From our internal use perspective, we decided from the get go to build and curate several mailing lists. Validating lists is a job in itself as any list builder will tell you. We could have bought in lists but we know from past experience that they will be riddled with error that are only discovered when yo use the bounce rates. What's the answer? We do it one case at a time. It is painstaking and time consuming but here is a smattering of recent results:
Open rates range 24-46 percent. Click through rates vary between 7-14 percent. In addition, we are seeing quantum jumps in both engagement and growth metrics. I am told those metrics are above industry averages. OK - so we're not trying to 'sell' anything per se and what we are providing is discretionary content where the recipient may well archive. We have seen that behavior as well.
The point is that we are eating our own dog food and learning what works and what doesn't. We won't get it right all the time and neither will you. But blind acceptance of automated methods of communication and marketing is not the way to go.
As a final word, Lauchlan participated in a webinar on this topic. It offers a pragmatic view of this topic. Check it out here. (Registration required.)