Marketing To Crack Whores With Kids

Chewy Chong's presentation last night, titled "You, Others and the Business of People." was a tour de force. It wasn't particularly cohesive, and I didn't really get what underlying message he was trying to leave us with - but it didn't matter. There were too many valuable ideas in his presentation to ignore. And I think the bit where he talked about making us think - well I stopped and smiled at that. I really, really did think during the presentation, and I've come to realize that I've missed that kind of thinking.

Chewy has thrown at us quite a number of valuable ideas. To make sense of them, I'm going to deal with the ideas in (more or less) chronological order.

Play to your strengths.
I first heard of this a year ago, when I was over at my grandparent's place and Oprah was on. This man came on stage to talk about how 'fixing flaws' and 'focusing on weaknesses' were the wrong ways to go about doing things. I'm not sure if there's an upper limit to the idea. In principle, at least, this makes sense: I feel impotent in my math classes, and empowered in CS3216, and I suppose this is because CS3216 is my kind of thing and math isn't. But does that mean, then, that I should give up on my math altogether? What if - like many other people I know - it just means that I am lazy, and that I am not willing to put in the kind of effort that math requires? Does that mean that I shouldn't choose to do Computer Science (which I, well, think is cool) and do IS/e-commerce instead?

People are not like you.
Laurence recently posted a link in his blog about a funny phenomenon where people, googling the term 'Facebook login', ended up at a blog about the Facebook login.

There seems to be increasing anecdotal evidence that people use Google as a kind of natural-language command-line interface. We geeks cannot imagine this, of course, but Google must be to these people some kind of tourist guide or god, one who would take them around a virtual world they do not understand. People do not think like you and me. We know this, on some level, but I don't think it ever hits us hard.

Many geeks are now flaming the iPad. They do not understand that Apple is not thinking about them. The iPad is a computing device made for my grandmother, my aunt, my technophobic sister. It is not made for geeks.

It is understandable that the geeks are not happy with this arrangement.

I must point out that there is one scenario where the principle of 'people are not like me' fails. If you are looking for an interesting problem to solve, for instance, it would be far more efficient to solve a problem that you have already have (as opposed to finding an hypothetical problem affecting someone else). Granted, the problem might be interesting but not marketable, or it might be that your solution to the problem is too unwieldy for the vast majority of people out there. But that is what iteration and user-testing is for.

I think I now understand why startup founders often endorse 'eating your own dogfood' and 'iterate fast and furious'. The first is true because you understand your problems best; the second is true because - as Chewy says - people are not like you.

There is value in CPA.
I never understood before this lecture what so many startups saw in online couponing systems. I believe I do so now.

Chewy's done a really good job presenting this as a challenging, interesting problem, with a pot of gold waiting at the end for whoever solves the icky thing. I've not given much thought to it, as he has, but there are some things about the problem that sound remarkably familiar. MacHeist, for instance, is a good example of a CPA marketing campaign. Is there something to their model? Or is it luck?

Regardless, if you'd like more thoughts on CPA, go to Shannon's blog, she's got some good thinking done on couponing systems in Singapore (and the like).

There is more money, currently in CPM.
Facebook is aiming for a portion of this advertising market. They believe that they can do what Google has failed to - attract big name, powerful brands who would pay hundreds of thousands of dollars for the digital equivalent of a glossy, full-paged ad.

Chewy is right in saying that CPA is an interesting, unsolved engineering problem. He has, however, neglected to mention that nobody has quite figured out how to make CPM work on the Internet. Facebook thinks they've got a solution - and maybe they do. But it would be interesting to imagine how one might go about attracting big name advertisers onto the platform. (And how about newspapers? Or sites outside the Facebook ecosystem. What solutions exist in that space, then?)

Is Bing worth it?
There is a signal to noise problem on the Internet. Bing believes it has a solution to this problem, and I wish them well. That said, I'd like to think about the implications of a machine doing information processing on behalf of a user.

First: is Bing a good idea? If we're talking about accuracy, Google has a five-year head start over Microsoft. It has also got better engineers (multiplied with less bureaucracy) working on the problem, piped in straight from Stanford. So it's unlikely that Bing's results would be as deep or as accurate as Google's are, at least not for the next couple of years.

Second, Bing's value proposition is to help users filter signal from noise. There are two ways to do this. The first way is to focus on the way information is presented. This means executing very strongly on design, and writing tools that convert information into visual representations of data. The second way is to make decisions on behalf of the user, in order to reduce the amount of information thrown at him or her.

The first option is something Google is not very good at (design). The second requires competing directly with Google's engineers.

The first option also hides from its users the fact that Bing isn't as accurate as compared to Google's searching algorithm. Humans, after all, absorb visuals better than they do text. And so, if Bing makes things very intuitive, it is likely to be more useful than even Google itself.

Which option is Bing currently doing?

But let's have some fun. Assuming that it takes the second option, Bing exists slightly higher up on the processing spectrum than Google, and this has its own set of problems. What problems might those be?

Well, how does Bing decide between two kinds of information? Bing's model implies that it would have to make certain decisions for its users, in order to abstract away some information that other search engines normally display.

So, say that Bing picks flights from a certain airline, for instance, over flights from another one, and users won't know because the raw information would be abstracted one level away from them. (This is a bad example, I know, but bear with me). What would this mean for Bing? What would this mean for Bing's users? I'm not too sure, myself, but I believe these are questions that we'll all have to think about. My current takeaway is that Bing would either have to be so incredibly useful that the users don't care, or that it would have to write really clever algorithms for 'information processing. But isn't that a conundrum: decision implies making one choice over another, and how would a machine know which information to choose?

[Quasi-related thought: what other solutions might there be, for this signal to noise problem?]

Conclusion
I enjoyed last night's lecture. There was this last bit at the end where everyone started talking about how Singaporeans aren't entrepreneurial enough, but I'm going to leave my thoughts on the issue for another day. Till then, there's the Wave assignment to worry over.

Signing out.