Hacker News new | past | comments | ask | show | jobs | submit login

Modern web advertising networks are an example of data telling lies we (or advertisers) want to hear.

The dream has always been to connect each advertising dollar spent to revenue generated. The reality is that data only exists in people's heads, is spread out over time, or is in their social interactions. My hunch is at least half, if not more, of all positive impact of ads (for the ad buyer) is generated this way. The data isn't low quality... it is literally impossible to collect.

For example: You research items on your desktop, then actually make the purchase on your phone while taking the train home. Maybe they can connect you to that purchase with enough tracking; So what if your neighbor or coworker asks about the product? Can't track that one.

Another example: You hear ads on a podcast for Igloo and when work starts asking about preferences for Confluence vs Jira you mention looking into Igloo. Your company ends up adopting it. Later you grow from 20 people to 3000. There's absolutely no way for Igloo to connect the dots leading to a 3000 user account. Let's say after discounts that's $300k/year. If Igloo paid $200/episode * 52 episodes/year * 20 podcasts = $208k. That's an absolute steal just to acquire that single customer.

Yet another example: You may be 24 with no kids living in an apartment but will you always be that way? Smart car makers understand that if you have a good experience with your first car in their brand you're much more likely to buy a larger car from them when you have kids, or a more luxurious car when your career advances. How can you figure out if the dollars spent advertising to the college kid with no money is wasted in that scenario?

So that's the great lie... that somehow all the tracking cookies, comScore profiles, etc will make advertising more effective or has some benefit period, regardless of fraud or click bots.

Advertising on the web right now is a fool's errand in many cases. I'm glad because it gives the small guys like me a chance to exploit the system to buy ads on lower-volume sites directly from the content creators and reach a good sized audience.




You can track all of that stuff, you just need enough data to be able to draw inference from the statistics.

Ultimately, if lots of poor people are worth targetting in a hope they will one day be rich, the statistics will ultimately back that plan.

But I suspect that anomalies will fall out in the big picture.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: