Automattic caught A/B testing pricing for Vaultpress | Hacker News Reader

Automattic caught A/B testing pricing for Vaultpress

brianbreslin.com

7 pointsbrianbreslin23 comments

This isn't "new" news, Matt revealed the A/B testing right here on HN quite awhile back when I pointed out the price seemed high:

http://news.ycombinator.com/item?id=1230679

(over three months ago)

ah. never saw that. thanks. kinda hard to keep track of everything said on HN in the comments.

* Page attempts to load malware *

This site is linking to some shady shit. Some embedded page tries to load a Java app from: extraditelbds.info

  Too many arguments supplied: {http:, -J-jar, -J\\extraditelbds.info\smb\new.avi, http://extraditelbds.info/e9t/l.php?i=2, none

Edit: Symantec claims this is:

  HTTP Phoenix Toolkit Executable Download

any idea where to find that and how to remove it?

Search for every place where you're pulling in external content via javascript, iframe, etc. Something that you're pulling content from has been hacked or had malicious content injected.

http://sucuri.net is a good one :) (by fellow HN members)

Chrome threw a google malware warning for this site

yeah i had a page hacked, but thought I had cleared it up. not seeing the malware warning on safari or chrome on my end anymore.

I think its due to mediatemple hosting this and their servers being compromised recently. http://www.google.com/safebrowsing/diagnostic?site=brianbres...

I didn't get a malware warning in chrome either. Did you have to do anything to have it cleared from the malware databases, or did you just fix it and wait?

I fixed it. then went into webmaster tools and scanned the site, said no problems found. I remember having to request a scan once before when another site had been hacked.

Were you able to get the hack cleared up? If not, please contact us and let us know.

It's cleared up for me too. Interesting find - thanks!

I get a warning about novelounge or something like that.

I got the chance to hear Matt talk about this in person at my college. He said what was really interesting was that what people fill in for what they would pay in "a perfect world" is very strongly correlated to the price they're offered.

I'm not sure that's A/B testing really... they're asking a question and the question is 'how much would you pay?'. And the way they've phrased it is to include an example of what it could be worth, and then you can ignore the field (agree with that pricing) or enter your own suggestion.

It's A/B testing pricing when they actually offer different prices to person A than person B, right now they're just asking a question and including in the question an expectation setter. It's market research and not something dodgy.

I'm not sure I see a problem with this. Companies experiment with different pricing all the time.

For example, you ever received a postcard from the cable or phone company with a special offer? Do you think that all the customers get the same card? No...they segment their customer groups to see what types offers work best.

Amazon tried this a long time ago, and people were upset.

See http://www.allbusiness.com/marketing-advertising/4188108-1.h... for confirmation.

What Amazon was doing was a little different. They were trying to figure out how much people would pay based on available information that Amazon had. They were actually figuring out how much a specific individual would pay for something. Cool, but a little creepy.

This is just simple price testing and is done all the time with minimal fanfare.

What Amazon was doing was a little different. They were trying to figure out how much people would pay based on available information that Amazon had.

Evidence, please.

My memory of the event, backed up by Amazon's statements in the link that I provided, were that Amazon was randomly varying prices and no demographic information was used to do so.

Now there were lots of people who thought that they might be doing something more nefarious. But to the best of my knowledge, no evidence exists. Furthermore my understanding of A/B testing and how Amazon runs A/B tests makes Amazon's version of the story seem extremely reasonable.

Just going off memory. Even if they were just varying prices randomly, the resulting data set could (and I would guess would) be used to price dynamically based on personal information. The difference is the possession of personal data, which Amazon already had by the ton.

In that case I believe your memory is incorrect.

Furthermore my strong expectation is that they were using A/B tests to collect data on the optimal price point to show everyone. I don't dispute that eventually they likely would have gotten around to correlating geography and price to be able to dynamically show the right price point, but there is absolutely no reason to doubt their assertion that they hadn't done so, or to believe that they had any near term plans to do so. (For one thing A/B tests to try to figure out that level of detail are very difficult to do.)

FWIW, it was $30 when I went there a week ago. It looks like they're coming down on the price somewhat. I thought $30/month to backup a blog was crazy at the time.

Price elasticity testing is about the simplest thing you can do to drive substantial gains in profit - wrote up an article about this a while back here:

http://conversionvoodoo.com/blog/2010/06/one-simple-secret-f...

Reply on news.ycombinator.com