Blake Borgeson, in blog form

suspected facts. validated opinions.

my comment on seth godin’s post “breakage”: be careful with your customers

with 4 comments

Seth Godin‘s blog doesn’t seem to allow comments, so here we are.

I really like the theme of his most recent post, “Breakage“: you have to be careful with your customers.  You might go years upping prices bit by bit, then all of a sudden lose half your customers before you know it.  Massive swings like that can happen whether making just a slight change to your pricing, or what you offer, or your tactics.  Here’s the last bit of the post: (although you should just read it–it’s short.)

When you hit the breaking point with one person, it might be 1,000 or 100,000 people who do the same thing at the same time. And you don’t get a second chance. They’re gone.

It’s not just money. It’s service. Or trust. Or spam.

You can stretch a rubber band for a long time. But then it breaks.

I heartily agree.  But he neglected to discuss what that company should be doing differently: it’s something so many businesses should be doing, and _especially_ businesses with a big customer base, since it’s so easy in the internet age.  Test.

Test new products internally, then with a limited set of your customers, before offering them to everyone.  Test advertising messaging, test landing pages, test changes to your checkout process if you’re in e-commerce.  And if you need to make pricing changes, absolutely test those changes before rolling them out to every customer.  Testing is the clear victor over focus groups, surveys, or any other method of gauging customer preferences without them actually making the decision in question.  If you ask them how a change in pricing would affect their decision, you won’t get the right answer.  If you serve 5% of your customers a different price, you’ll quickly know if it’s a good or terrible idea.

Maybe Seth’s insurance company, in fact, was merely testing pricing changes, and he happened to be in the 5% of their customers randomly selected to get the highest price increases, year after year.  If so, then Seth’s departure has just given them some very valuable information they can now apply to their pricing model for their customer base as a whole.  My gut tells me Seth’s right, though, and they’re just raising prices for everyone by the same amount, year after year.  If so, then as he suggests, they may have made an unexpectedly costly mistake.

p.s.: There are plenty of businesses where testing prices often doesn’t make a lot of sense, for a number of possible reasons.  At basecamp, for example, simple and consistent pricing is part of what makes them who they are.  I wouldn’t favor pricing experimentation for them.  For amazon, definitely yes.  For a new startup still looking to hit its product/market fit: don’t worry about testing tiny changes–iterate on your whole product.

Written by blakeweb

September 13, 2008 at 3:47 pm

Posted in marketing, startups

Tagged with , ,

4 Responses

Subscribe to comments with RSS.

  1. I read that post yesterday as well and decided that I disagreed, in general, with his assertion that you can't treat supply and demand as a curve. I think that there are probably very few cases where you raise prices by 5% or 10% and you lose a substantial chunk of your customer base forever. Even if it were true that Seth's insurance company lost a substantial chunk of customers during that price increase, it's probably because the Insurance company jumped too far up the curve where the curve looks more exponential. That would be a result of anticipating the curve incorrectly — and as you point out in your 'comment,' testing is a great way to help define the curve so you know what to expect with price changes.I do think it could happen with trust more easily; with the most extreme cases being a bank that skimps on some software and ends up constantly giving all their customers' money away to someone else. Although car companies kill hundreds of people a year due to shotty design and they are still in business — so who knows?

    Dan Graham

    September 14, 2008 at 2:33 pm

  2. I'd love to read some of your ideas on how to test pricing on your customers. I really hadn't thought of it. But it absolutely makes sense. Any resources you come across in terms of price testing would be great, too.


    January 12, 2009 at 8:11 am

  3. Hey Jonathan–great to see you here!My favorite resource for discussions about versioning so far is Eric Ries, co-founder of imvu. He's incredibly smart both technically and business-wise and just started writing a fantastic blog about 4 months ago at I imagine you've seen links to some of his posts on hacker news. He's very prolific as well. To thin it out a bit, here's a link to a tag of split-test posts he's written:…He talks about doing all the versioning the manual way deep within the code for the site, which is how we set it up as well. But you can use some tricks to leverage google's website optimizer if you're limited in your coding bandwidth. For example, if you've got an out-of-the-box ecommerce solution or you coded one yourself, just run some different versions of portions of your homepage that link through to duplicates of the same product, but at different prices. Then google handles all the cookies, tracking, and conversion analysis. You'll have to do some math yourself to pick the optimal price for you based on your profit margins–conversion rates is, I believe, as far as website optimizer will take you. Hope that helps!


    January 12, 2009 at 8:35 am

  4. Thanks! I'd looked briefly at Google's website optimizer. I'll take a look at Eric's posts, too. I just found the blog from your profile on Hackers and Founders. Good stuff. Keep it coming.


    January 12, 2009 at 4:21 pm

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: