Good analytics, strange application, and unintended consequences

Over the last few weeks, both at a professional and a personal level, I have been struck by how analytics, which should drive sensible behaviour, which is clearly justifiable and logical is not achieving that aim.

One of the projects I am working on is using an excellent real time operational intelligence tool that should be able to send highly targeted messages to the right people at the right time; but the test results from an initial trial are proving to be less positive than hoped for, not least because far too many of the selected individuals cannot be contacted because they have opted out of all marketing communications. The reason that they have opted out of marketing communications is that even with a data warehouse and advanced analytics providing a basis for making highly targeted campaigns, there is still a reluctance to just target the high probability candidates, and the long held belief from the days of mass marketing that there is a need to have as big a funnel of prospects as possible still persists. So that instead of only contacting those who have the greatest likelihood of responding favourably, we are still seeing a mass contact strategy that people find irrelevant and, understandably, they would rather hear nothing. So despite all of the capability that should exist we are still spreading a vaguely targeted message at a vague likely to respond set of prospects, which is not what was ever intended.

Then I had a really interesting interaction with online security analytics. My eight year old Mac started to get very unpredictable so I decided to buy a new one. I went online, used a credit card linked to the store, ordered it to be delivered to the store nearest my home, and had the order confirmed. Then I received an email which told me my order was declined as I had failed a security check. I was first told it must be my credit card but when I contacted the card supplier I was told that, no, they had authorised the payment, but the retailer had cancelled the transaction. I then contacted the retailer again to be told it was something to do with the location from where I had ordered it, which was from a head office building in Slough. I then reordered and had the goods delivered, but I asked for an explanation. The response I got was to be told they could not explain why a valid credit card, from a valid ip address, asking for delivery to a valid address should have been refused, because it was compromised by their security checks. At no point was I at any stage assured that no record or action will be taken as a result of the refusal and that no information about this refusal will be shared with any other party. As a result I doubt that I shall be using these retailers’ online services again, and will be looking closely to see if I get negative responses from any other bodies. They may think that they are never knowingly undersold; but I feel very seriously underwhelmed by what they have done.

I would be interested to know if others have examples of security and marketing analytics resulting in similarly negative outcomes. This is not what the analytics community need if consumers are to have confidence in a digital future.