My partner Nikki Baird and I saw a lot of interesting things last week at the Internet Retailer show in Chicago, but chief among them was Jason Goldberg’s presentation on opening day. President of Target.com and Mobile, Goldberg came to Target in 2013, after having spent years at Gilt, Hayneedle, and Amazon.

Whereas a lot of presenters tend to keep a keynote session pretty breezy, the first thing Goldberg did was address the Lilly Pulitzer fiasco from six weeks ago. And both the sources of that error – as well as the lesson to come out of it – may surprise you.

As you likely recall, part of the problem back on April 19 (when Target began selling a 250 sku, low-cost collection of Lilly Pulitzer-designed apparel, accessories, and home décor items) wasn’t just that inventory ran out rather, ahem, quickly. It was also that digital demand for the Lilly line crashed every one of the company’s digital offerings. While a lot of different answers have come out of Minneapolis on how long the company really expected the inventory to last (their spokepeoples’ initial response was that the retailer expected stock to last for weeks, something few of us took seriously), Goldberg didn’t mince words. “I knew we’d sell out:  maybe in hours… maybe in minutes. ” But from a tech standpoint – his domain – there wasn’t going to be a problem.

The promotion was slated to begin Sunday morning at 6AM. Goldberg went to bed confident on Saturday night (more on that in a moment), so much so that he debated whether he’d even need to go into the office on Sunday, but elected to in order to be part of the fun. He already had the subject line in mind for the email he’d craft the next day to Brian Cornell, Target’s CEO: “Huge success. ” Then all hell broke loose.

The site went down, and he and his team had no idea why. “For each 10 minutes that passed we were always 10 minutes away from figuring it out. ” The subject line of his email had to change from “huge success ” to “Brian, we have a problem. ”

In the end, it turns out that what crashed Target.com was the one thing Goldberg’s team didn’t test for. Previously unseen numbers of shoppers were trying to conclude purchases on the brand’s mobile app. When it couldn’t handle the volume, it took everything down with it. Now here’s where this gets particularly interesting. Goldberg openly admitted that he was confident in the technology’s ability to handle the promotion with ease. And he had a right to be. His team had “tested everything ” based on data Target had collected just six months ago from its Black Friday promotions. They’d handled Black Friday beautifully, and this promotion wasn’t even going to come close to that kind of volume. By all accounts, the company had been “on fire ” with their digital offerings for quite some time.

But if we learn anything from this story it’s that even data from as recent as a few months ago can’t tell us how customers shop today. In his own words, “We missed the mark on Lily because we relied on customer data that was six months old, and even that wasn’t good enough. ”  If you stop and think about that, it’s actually kind of mind-blowing. In such a short window of time enough people had changed the way they shop – from mobile exploration to a desire to take fulfillment via mobile app – to bring a Fortune 50 company to a screeching halt.

I’m really glad Mr. Goldberg took the opportunity to tell his story. But at its core, apart from serving as the rare and honest look into a giant retailer’s inner workings, and a lesson in how much faster an enterprise can learn from mistakes when accountability is applied, it has to make us all wonder: what are some things we all might be overly confident in today based on our ability to deliver yesterday?