Conversational marketing – will it scale

By August 7, 2007From mobile

Posted by mobile phone:
I`on my phone here because I can`t get the wifi at Aspen airport to work – so this will be a short one with no links.

I attended a panel session today with a guy who helps Nike and Starbucks, amongst others, with how they should approach marketing on social networks.

The good news is they are keen to do things. They want to spend money to reach 15-30 year olds and this is an obvious place to do it.

I am pretty well convinced that the best way to do this over the longer term is conversational marketing. As I`ve said before people are getting increasingly blind to banners. That was a sentiment echoed at the conference. Putting video into the banner will help, but I get the sense that most people only expect this to have a temporary effect on banner performance.

The question I have been asking for a while now is `how will conversational marketing scale?`

I asked that again today and the short answer is that nobody knows. Not yet anyway.

Doing conversational marketing well requires consistent and ongoing effort from the agency/brand. Compared to most forms of advertising the cost is relatively low – but it needs a lot of man hours. That is the scaling challenge, in a nutshell.

One thing I hadn`t realised until today is that conversational marketing doesn`t necessarily translate into big bucks for the social net. Brands are not so much buying media from them as participating in the community. That said Myspace is apparently now charging six figures to set up a profile for a brand and seven figures to market to it.

  • Martin Willitts

    I have a suggestion for making it scale. Using a piece of technology called a lingubot provides the means to create a sense of dialogue. Currently used by the likes of Ikea and Lloyds TSB, the visitor/customer interacts using natural language and the responses are driven by the technology. I would love to tell the “guy who helps Nike and Starbucks” more about this.

  • Martin Willitts

    I have a suggestion for making it scale. Using a piece of technology called a lingubot provides the means to create a sense of dialogue. Currently used by the likes of Ikea and Lloyds TSB, the visitor/customer interacts using natural language and the responses are driven by the technology. I would love to tell the “guy who helps Nike and Starbucks” more about this.

  • nic

    It would have to be some pretty special software Martin. The language and dialogue would need to be much more varied than for customer service apps.

  • nic

    It would have to be some pretty special software Martin. The language and dialogue would need to be much more varied than for customer service apps.

  • Martin Willitts

    http://www.elbot.com/ and http://www.yhaken.com/bot.php are two examples. Also, the technology’s been used for the BBC’s virtual character Jamie Kane (http://www.bbc.co.uk/jamiekane/). Furthermore, as demand for particular conversation topics reaches a certain threshold, support for that topic can be added in readily.

  • Martin Willitts

    http://www.elbot.com/ and http://www.yhaken.com/bot.php are two examples. Also, the technology’s been used for the BBC’s virtual character Jamie Kane (http://www.bbc.co.uk/jamiekane/). Furthermore, as demand for particular conversation topics reaches a certain threshold, support for that topic can be added in readily.

  • nic

    Martin – both these demos are quite impressive from a computer science point of view and I could buy into the idea that they might be able to handle some customer service functions. They are a long way from sounding like humans though – and this is the key for automating engagement marketing – your bots would need to pass the Turing test. It is a bit early to be saying that categorically, but I suspect that will be the outcome.

  • nic

    Martin – both these demos are quite impressive from a computer science point of view and I could buy into the idea that they might be able to handle some customer service functions. They are a long way from sounding like humans though – and this is the key for automating engagement marketing – your bots would need to pass the Turing test. It is a bit early to be saying that categorically, but I suspect that will be the outcome.

  • Martin Willitts

    Perhaps there’s a point there. By fundamently being difficult to scale the human interaction, a different problem should be tackled. Converting the challenge to being maintaining some interaction with customers there is scope for introducing a fully declared non-human ‘character’. This could make a virtue of the situation and the nature of the automation would be adapted to fit the brand.

  • Martin Willitts

    Perhaps there’s a point there. By fundamently being difficult to scale the human interaction, a different problem should be tackled. Converting the challenge to being maintaining some interaction with customers there is scope for introducing a fully declared non-human ‘character’. This could make a virtue of the situation and the nature of the automation would be adapted to fit the brand.