The 7-11 Milk Experiment: How Does Site Design Affect Revenue?

Christine Perfetti

September 13th, 2005

At UIE, we’ve spent a great deal of time trying to assess how elements of a site design affect revenue. In all of our recent studies of e-commerce sites, we’ve based our research on what we refer to as the 7-11 Milk Experiment.

What is a 7-11 Milk Experiment? Here’s the scenario: Imagine I had a way to identify when someone has run out of milk. I pick them up in my car and drive them to the nearest 7-11. And just to make sure everything goes well, I give them the money to buy milk. How likely is it that that store can sell milk in this scenario? When I’ve asked people this question, almost all respond that it’s close to 100%. If the store does not end up selling to the user, there’s clearly something wrong with their experience on the site.

In our research, we mirror this type of experiment online. We find people who need products, bring them to sites that have the products they want, and give them money to buy the products. What did we find?

In our latest study, users only purchased 30% of the time! So, what was happening here? We found that on most of the sites, users just couldn’t find what they were looking for and that the site’s organization was to blame.

From watching users shop, we’ve seen that they use a progressive process. Users move from one stage to the next, as they try to purchase a product. One of the biggest priorities in our current research agenda is to identify where in the purchase process users fail, such as the Home Page, the Product Lists pages, or the Target Content pages. By understanding how these different stages and types of web pages work, we hope to learn a lot more about building usable sites.

13 Responses to “The 7-11 Milk Experiment: How Does Site Design Affect Revenue?”

  1. Eric Scheid Says:

    What were the other reasons users didn’t purchase? I can think of two: (1) loss of trust in that particular retailer being able to deliver the goods as advertised, and (2) deciding they didn’t actually need the thing they thought they did, once they got their hands on a better description/specs/photo. Were those reasons observed during your tests, and to what extent?

    Do you consider appearance of trust and compelling sales copy to be part of the usability problem?

  2. Antoni Dol Says:

    I guess you guys have read Paco Underhill’s Why we buy?

  3. Enric Naval Says:

    I have seen lack of trust as the first reason for people to refuse to even read the content in a site.

    Specially when they expect a certain way of doing things, and they get a different way. So, if you promise following standards, and then they see that your web relays heavily in table layout, people will bail out of the site without clicking on anything and they will badmouth you on their favourite forum, because they are so pissed off.

    Beware of any expectations you may be creating, purposedly or unadvertingly, because if you break those expectations then people will be very angry to you.

    Yes, they will be angry to you even if it is not your fault. Sad but true.

  4. Christine Perfetti Says:

    Eric: Trust did come into play for some of our users online. For example, one of the sites we asked users to visit was Many users didn’t want to buy from Walmart because of trust issues. Because of this, we actually had to remove from the study. While trust was a factor that caused some users not to purchase, it wasn’t one of the major reasons people left sites without buying.

    The two main reasons why users didn’t purchase in our study were that 1) they couldn’t find the content because of problems with the site organization and 2) once they found the page with their target content, the site didn’t provide enough information for users to make a purchase decision. In these cases, users decided not to purchase because they couldn’t decide whether the products were good enough. (Jared has written an article, The Customer Sieve about many of the different problems users encountered during the purchase process.)

    Antoni: We have read Paco Underhill’s book, Why We Buy. It’s a great resource that we recommend to clients to help them better understand shopping behavior.

    But because Paco Underhill’s research focused primarily on shopping in the physical world, it’s a little light on the topic of online purchase behavior.

  5. will Says:

    What about the shirtless gang of boys on their dirtbikes that hang out in front of the 7-11 and drink massive amounts of Dr. Pepper? These kids usually prevented me from going indside, as they intimidated this skinny geek.

    Equivalent to people being scared of phishing and other forms of theft perhaps?

  6. Ano Nymous Says:

    What is a “7-11”? Some kind of special real word store? Please, explain this to your international users who have never been to the USA.

  7. monk.e.boy Says:

    I think 7-11 is like 24-7, only it’s open 7 hours a day for 11 days a month. They only sell 7-UP to americans, so don’t worry too much about it.

  8. Jared Spool Says:

    The monk.e.boy is correct. 7-Eleven is probably the largest chain of 24-hour convenience stores. Read about ’em here.

  9. Enric Naval Says:

    I believe that the problems that you mention are causing the user trust to be lowered. When the user’s trust is low, they will give up sooner and/or fail more often when buying in your site.

    This is not a problem for you, but I’m sure that Wallmart considers that to be a problem 🙂

    By trust, I meant this sequence of actions:

    user enters your site with normal trust
    user can’t find a product that very clearly HAS to be there
    user lowers his trust on your site
    next time,

    user has low trust on your site, and decides to go somewhere else, raising its trust in a different site
    or, gives up sooner and lowers his trust even more

    I have seen that users who trust you will be motivated and forgiving of errors.

    So, users that don’t trust wallmart have already gone through this cycle several times. Now, they will even refuse to TEST the site, let alone buy on it.

  10. ae Says:

    You’ll forgive me, but I’m detecting a bit of political bias. TRUST, as relates to web-based retail shopping, would be expected to be quite high for Wal-Mart, a known name, with numerous hard-world addresses, and demonstrated accountability in terms of quality and delivery. Contrast this to Joe’s Web-Based Shopping dot com, which may not even have his actual hard-world address anywhere on the site.

    Really, what do you mean by TRUST? Not “trusting” Wal-Mart because it’s a giant mega-corp that is squeezing small retailers out of the neighborhood everywhere it goes is a very different concept than the sort of “trust” you’re referencing here in terms of market research. Please clarify. This murkiness is affecting how much I trust your observations.

  11. Ellie Says:

    ae – amen. this is precisely the reason many people refuse to buy from WalMart whether online or off. I don’t know whether labling ‘virulent hatred of a corporate parasite’ as ‘trust issues’ was a political choice on UIE’s behalf, but it wouldn’t surprise me.

  12. Missing the point « World of Usability Says:

    […] to online stores who offer that product and the results are only a 30% buy rate (i.e., the 7-11 milk experiment) , something is desparately wrong and it needs […]

  13. Compelled Shopping Test Says:

    […] is UIE’s description of the 7-11 experiment. This entry was posted in Uncategorized. Bookmark the permalink. Post a comment or leave a […]

Add a Comment