Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Using A/B testing to find story ideas (andymboyle.com)
9 points by danso on Oct 28, 2011 | hide | past | favorite | 8 comments


So basically an in-house, impromptu trend analyzer a la Demand Media?


Please do not do this. Everyone has a tendency to write about whatever's in the news today anyway--we will spot your over-topical posts a mile off.

Have an interesting point of view, or some experience your audience lacks. Be helpful, be opinionated. If you don't care about the topic why should we?

This is the second article on the homepage to mention adcenter, which I had not heard of until today. HN reader/writers: I'd be interested in an adwords comparison that did not boil down to "less eyeballs, cheaper."


I don't necessary disagree with you but there is middle ground here. While there is a problem with mass media only appealing to the masses (I.e. most of tv news), there's an opposite problem of an ivory tower editor caring only what happens in his or her world. there's an old saying in newsrooms that news is whatever happens to the editor.

So this technique could be used to suss out important topics that are outside of your newsrooms social sphere. And en evaluate the demand versus what your news outlet has currently produced for said topic.

For example, most editors live in their city centers and have little clue about the much more populous surrounding areas. So, not "Paris Hilton" and "Kim Kardashian"' but "Morningside Heights" vs "Flushing"


Fair point.

Given that you have to enumerate the topics ahead of time to create an A/B test, I'm not sure how this can help editors see outside their city center.

Personally, I don't read newspapers and greatly enjoy blogs with tiny ivory towers over which the author has a large amount of experience/insight. I can see that that may be a uniquely "internet kid" point of view.


The point I was trying to make is when editors have to make choices between areas to cover. As you may be aware, most traditional media outlets are becoming smaller in terms of news gathering employees. So now they sometimes need to make decisions about what areas they should be covering.

If less folks search and clicks on ads for "Cleveland education news" rather than "Cleveland crime news," one could attempt to use that data in making their decision if you had to choose between those two topics to cover. You're serving your readers and getting them information they want, rather than spending time covering a topic that won't be read by your readers. I think this would be a bad financial decision and a bad allocation of resources.

I guess I didn't make this point clearly in my quick blog post, but I'm not advocating for doing a Demand Media-style operation. What I'm advocating for is using more analytics to make your decisions in allocating diminishing resources. I think you should do that rather than just going by your gut.

As a person who's worked in the news industry for all of my short career, its tendency has been to make decisions not necessarily based on data. I'd like to change that.


I guess I'm confused as to what you're disappointed about. My audience is mostly for newspaper/media folk, and I tend to try and bring ideas from the web development world to them. So that's what I did with this quick post.

I'm not sure how I don't care about the topic, as I did write about it and offered ideas for how newsrooms could use this.

I hadn't heard of adCenter until today, either. But I thought it worthwhile to mention a tool that isn't Google for once.


The point is that article that newspaper folk produce using this method will essentially be along the lines of the Google-chasing Demand Media content farm stuff. Give people exactly what they're looking for at the top of their mind and a lot of important reporting goes away.

It's not your article he's complaining about, it's the articles that your article encourages people to write.


I'm not disappointed about anything.

You're on the front page of HN today, so whatever your regular audience, your audience today includes a number of tech bloggers.

For newspapers: you will not be catering to your audience, you will be catering to the people who clicked on those ads.

I was not suggesting you did not care about your topic.

I was not suggesting it was in any way improper for you to mention adcenter. But it's funny, you just parroted what was in the news and instead of useful trend data about the popularity of adcenter among bloggers we just have more noise.




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

Search: