The flawed optimism of digital advertising models

NewImage

“Do you do any work on how annoying you are?” – Peter Day to an ad re-targeter…

In Business, the podcast by the BBC’s Peter Day, is something I have enjoyed for years. Every now and again he does a programme which is so exactly pertinent to things I’m working on that I listen to the whole thing with a broad grin, while a sort of Hallelujah chorus jangles about in the back of my mind while I am listening…

For Your Information was one of those. I’ll be coming back to some of the trains of thought that departed this particular station in some future blog posts, especially its focus on information overload’s effects on productivity and organisational effectiveness, which connects directly with the web superskills theme I talked about at TEDx in January.

For now, the thought I want to share is one sparked by a comment that Peter Day made to a online advertising re-targeter: “Do you do any work on how annoying you are?”

He was talking about the irritation he might feel when an ad for something he searched for days ago followed him around for days afterward.

The response wasn’t convincing. Of course you don’t want to annoy your potential customers, said the re-targeter, and they provide tools to help you not blitz people.

I wonder how many users of the system calibrate in that way?

Re-targeting is just the latest in a long line of advertising technologies and innovations – latterly mostly in digital – which promise – and often deliver – “uplift”, greater click-throughs, sales, awareness etc. than previous methods.

There are two broad responses to an ad following you around the web. The first is “Wow, that’s cool!” The second is a raised eyebrow, a suspicious sneer, a question: “Why is that happening? How do they know who I am? What elese do they know?”

Response one typically comes from, er, people in digital advertising. The second, in my experience, comes from anyone else.

The steady flow of privacy nightmare stories and Facebookphobia in the media and generally in people’s consciousness is raising a – probably healthy – scepticism about online media. The more digitally literate the average user becomes the more they question what is happening to their personal data and how it is being used.

Wishful thinking on the part of online media companies and digital agencies means that not enough work is going into thinking about this growing, fundamental user need.

Apart from the inconvenience of facing up to the possibility that people might not want to play exactly the role alloted for them in the great media/marketing ecosystem, the digital advertising industry is let down by a kind of fatal optimism. They only want to look at the good news in the data and not the bad.

The thing is, that the bad news might be as useful, even more useful than the bad.

To illustrate, take a look at “success” in a typical online display campaign. A clickthrough rate of  0.2%.

Doesn’t matter what happened to the other 99.8%, i.e. most people. They simply weren’t interessted enough to look.

When it comes to re-targeted ads, social ads, etc., the clickthrough rate improves over that of typical ads. I wonder if annoyance and negative feelings to brands using these techniques does too?

Clues about what people don’t like in ads - signals of dissatisfaction, if you like – abound, but it always the positive outcome on which paid media professionals are focused. Maybe there would be more use in looking at all the data, including the damage you may be doing your own cause?

 

 

 

Facebook planet

201101101112.jpg

This image has been on my computer desktop and on my mind since I saw it in December. High time I shared it here, really.

It’s a data visualisation of 10 million pairs of friends on Facebook and where they live in relationship to one another, created by an intern on Facebook’s data infrastructure engineering team. Read the original blog post in full – it is fascinating stuff.

As Ian Tait points out, what’s amazing is that there is no map underneath, and yet you can pick out the shapes of the continents.

Interesting too are the gaps – China, Brazil and Russia are underrepresented, perhaps due to the fact that other social networks are more prevalent in those territories (RenRen, Orkut and Vkontakte respectivelY).

Via Broadstuff. Reminded by Crackunit

Customers in networks: slides and notes from ChangePlayBusinesss

ChangePlayBusiness was an unusual event, to say the least, living up to its promise to be an unconference. About 40 innovators and entrepreneurs gathered at the ICA to play a game about creating businesses, the playing of which included connecting with one another (there were a lot of interesting people) and meeting subject matter experts on everything from financing to marketing (which is where I came in).

My role was to deliver a “masterclass” on understanding and communicating with customers in a “changing economy”. I chose to interpret this as an opportunity to talk about businesses in the age of networks, in the age of complexity.

The slides are here for those (with the push/pull error reversed!) who attended the session:

And these are the links that I promised to post:

Mining the where: CIA-funded start-up says 80% of online content contains location information

201011251323.jpg

Last week at Local Social Summit I talked about some of the issues around location and privacy. Especially problematic is people’s inadvertent tagging of their photos with location information – something potentially of interest to crooks, stalkers and others you might not want to know everything about you.

Open-source spying is a term which has been around for a while, reflecting the fact that when it comes to gathering information, the web is often as good a place as going into the field. In-Q-Tel’s investments reflect a justified fascination with the social web by intelligence agencies.

Well it turns out the CIA is also interested in this kind of information. In a post about the CIA’s Silicon Valley VC firm, In-Q-Tel, the Not So Private Parts blog on Forbes found the firm…

…likes companies coming up with better ways to mine social networking sites and geospatial location data. One of its investments, Geosemble, a private spin-off from USC, estimates that “80% of online content has location information.”

80%? Wow.

“Our mission is to shine a torchlight on geographic unknowns and help organizations neutralize threats and capitalize on opportunities in their areas of geographic interest,” says its website. Another of IQT’s geospatial investments, FortiusOne, promises instant maps based on Tweets and photo uploads, for mapping election-day threats in Afghanistan, for example.

The idealist in me is attracted to the data mining stories of humanitarian efforts of platforms like Ushahidi, but we should remember that governments and their agencies are interested in our geo-location information as well.

Bonus link : : Really interesting presentation from FortiusOne on analysing geo-data for business.

Mining the now: Real time sense making

Here’s an emerging idea about how we can and will make sense of the vast amounts of data becoming available about what is happening right now…

Juliana Rotich (whose work at Ushahidi I’ve discussed on this blog about a year ago) writes on her Afromusing blog about pulling together the ability to crowdsource the gathering and making sense of data in real time (which is what Ushahidi and the related Swift River project do) with the concept of the internet of things (objects and systems connected to the internet and making data available continuously):

Allow me to use the word holistic. As in holistic near ‘real time sense-making‘, incorporating the internet of things, with crowdsourced data delivered through channels that encourage participation. There is an opportunity to see things dynamically and not just do after-the-fact post mortem. This could work for flash point events like the Haiti earthquake (taking data [from] Geiger counters etc + crowdsourced data like that available on the haiti deployment run by Noula.ht. It could also work for longer term events such as the BP Oil spill in Louisiana.

Juliana quotes Jeff Jonas, a scientist at IBM:

…the closer to real-time one can get the right answer and respond, the better. And milliseconds matter.

The concept of real time sense making offers so many tantalising possibilities, from predicting the behaviour of human social networks to helping those networks (countries, companies, NGOs) respond to emergencies and more broadly to the challenges we face globally.

Be sure to check out Juliana’s original post for some amazing examples of how this concept is being put to work.

201011251225.jpg

Image: Literally, a swift river…