StorefrontBacktalk: New Target.com Forgot the Customers
Through a special arrangement, presented here for discussion is a summary of a current article from StorefrontBacktalk, a site tracking retail technology, e-commerce and mobile commerce.
When Target’s development team launched the new Target.com on August 23 after two years of development, it must have felt like a dream come true. But it wasn’t — at least not for customers who discovered that big chunks of the new site didn’t work at all, and almost nothing worked as well as the site they’d seen a day or two before.
Why? What went wrong? Actually, not much, from the point of view of experienced developers. Naturally the site had glitches — that’s to be expected.
Target decided to end its arrangement with Amazon two years ago — and that meant it had a completely blank slate to start from in creating a new site. Most e-commerce execs would love that opportunity to shed all the legacy code, the decade or so of kludges, workarounds and hacks that make it so difficult to do anything really innovative. All that old junk makes new approaches next to impossible.
But the downside is that there was also no legacy code that worked. If a new feature was too buggy, there was no old version to fall back on.
Target’s developers figured that was OK. The site would go live, they’d work the kinks out as quickly as possible, soon there would be all sorts of great new stuff built on the wonderful infrastructure that was still invisible on opening day, and everyone would understand — right?
No. Customers neither knew nor cared that the new website was the product of two years of loving development and was bound to have a few hiccups at first. It didn’t matter to them that Target had to build from scratch or that all sorts of wonderful new features would be coming once the site was stable.
All that customers saw was that their passwords, which worked fine on Monday, didn’t work on Tuesday. They couldn’t edit their wedding registry lists. They could no longer track orders they had paid for a day or two before. The weekly newspaper ad wasn’t showing up; neither were coupons. A large digital countdown clock on the homepage (an extremely long homepage) warned that today’s Daily Deals would end in so many hours, minutes and seconds — but the link went nowhere.
In fact, lots of the links were dead ends, delivering customers to very pretty error pages featuring Target’s mascot dog. (There’s a downside to using pictures of a dog mascot all over your site, including error pages: at a certain point, customers are likely to start really hating the sight of that little dog.)
No doubt all of that will soon be fixed. Much of the site was working far better the following day. But it will take a lot longer before customers feel like the new site is as good as the old one — which, of course, means better than the old one.
Discussion Questions
Discussion Questions: Do you think Target bungled its new website launch? What’s the best way to prepare and reassure customers for likely problems encountered as part of such overhauls?
Glitches to be expected? A glitch is a fancy IT term for failure. So that means Target expected to fail? The distraction factor is so big when it comes to online shopping, retailers cannot afford ‘glitches’ in their site. Especially a big brand like Target. This site should have been beta tested with higher value customers first before anything was launched publicly. And if I was a Target executive and some pocket-protector-wearing IT guy said ‘glitches are expected’, I would have a new IT guy within the hour.
Simple answer: Yes, they botched the launch. In fact, when I went to target.com before writing this comment, I was greeted by the error message and the picture of the Target dog. This is amateurish, especially given two years to get it right, and it’s hard to judge the workmanship of the site if you can’t even get inside! (My wife has used the new site and commented that it is more difficult to navigate than before.) There may have been been issues with the Amazon partnership, but the new site should have been launched after extensive beta testing in parallel with the original site.
I’m flashing on countless Dilbert cartoons here. Target’s website, IMHO, was never that good — lots of search engine issues, but to put out a site that doesn’t function properly is going to drive people away. It’s not like we have to shop at Target.com.
Using the top banner to warn users and maybe even enlist their aid in reporting problems might have been a very proactive move.
Note that as of 9:42 am today, the website is down — just a picture of the dog to tell us they’re working on the site.
If passwords didn’t work and bridal registries could not be accessed, Target bungled their new website launch. They should have warned consumers well in advance that a new site would be launching and recruited them to help make the site better by finding glitches and reporting them.
In advance of the consumer launch, Target should have made the site available to a team of experts (they are easily hired), and let that outside firm pound away on the site.
In 2011, there is little excuse for launching a site with significant problems.
Target’s inefficient launch of its new website reminds me of the Viking’s first play on offense last Sunday against the San Diego Chargers, which was an audible, an interception on a short pass and a quick touchdown for the Chargers.
Two years of preparation should have prepared Target to initiate its new website without those glitches. What happened caused some first impressions of communication impotency.
Customers and potential customers don’t really care about Target’s website problems. It would seem those two years of “spring training” didn’t produce a good game plan. That seems so “UnTarget.”
For some reason the online development world does not believe in testing. Would a company accept a new accounting application that had bug in it? The answer is no. When I was programming, the company had a tester for all new applications. If the tester could break the program, it was not ready for prime time. A contributing factor for online development is the fast pace of change. Anyone that spends 2 years developing a web site will either be obsolete by the time they are finished and have numerous problems to every changing design. Too often developers are into the wow factors not functionality. A better approach is to get the base in place and continually add.
Retailers and manufacturers sometimes forget that consumers have lots of choices and that their loyalty to a certain retailer or brand is only as solid as the user-friendliness and predictability of the retailer or the brand itself.
Well, I was about to launch into a whole diatribe about how people don’t appreciate how hard it is to build a website with tens of thousands of new products in two years, with ZERO legacy code (including pictures, I suspect). But then I read Stephen Needel’s comment that the site is down. Sure enough, 10:23 am and the site is still down. That’s not good. And it’s really ironic.
Questions abound: How many programmers were working on the site? Where is the site co-located? What’s the redundancy plan? Contingency plan? Of course, since RSR’s site was down for about 5 hours a few weeks ago because of a faulty switch at the co-lo facility (a very large one, I might add) I can’t throw a lot of stones here. Still, we expect glitches when a site first comes up — that’s why you bring them up at low-volume times of year…but it has been 3 weeks now. Time to tidy the thing up.
None of us are naive enough to think it’s easy (at least I hope not). But sometimes I think we’ve lost the basics of sound systems design and contingency planning. I call it “The Facebook Effect.” Facebook just sort of changes things. No warning, no notice. Just something new. Sometimes the change works. Sometimes, not so much. But if 750 million subscribers don’t mind (and they tend not to), that attitude slips into eCommerce app development as well. Too bad, really.
Moments before posting this comment I checked the Target website and the entire site is now under construction, unprecedented for a company of Target’s reputation.
The best way to manage customers expectations is to exceed them. Target should have identified potential problems and built in solutions before launch through a robust testing of the new site. Additionally they should have built in redundancy and maintained their clients data on the old site in the event that major problems required them to take down the new site.
Parallel programing could have avoided most of the major issues while leaving Target with fewer and more minor customer issues to deal with.
There’s an apt analogy for this. The approach shown by the Target team is very similar to Microsoft — hey, we’ll learn what doesn’t work by the complaints we get when we launch — vs. Apple, who starts from the desired experience of the end user and works backwards through the source code. The first approach inspires hackers and outraged customers. The second produces between the 1st and 2nd highest market cap in the world.
This just in…the twittersphere reports that Target’s site is down because of crazy traffic when it launched the Missoni line this am. I think they call that “stress testing”….
Wow, the site is down today. I suspect that the IT person in charge is a little worried about his job security.
I’ve always loved the way Target approached the retail experience as compared to other large retail stores, but I think they really missed the boat on this re-launch. I was impressed with the response of the site when I tried to log into my registry though. I received an email from target.com for 20% off any online order for the inconvenience. But this seems like a high price to pay for poor planning.
Just a quick followup comment: Two hours after my first post, and the site is still down. This time, it’s because Target is “extremely popular.” Could have something to do with today’s launch of heavily advertised Missoni merchandise. Embarrassing…what happens on Black Friday?
How embarrassing and costly for the Target folks and disappointing for the company’s loyal shoppers. We may expect some heads to roll over this. High-up heads.
I’m no expert on e-commerce site development, but after a decade and a half of commercial experience, I know there are pros out there with the required know-how. I’m pretty sure they would not recommend a ready-fire-aim development process.
All retailers should take a sober look at this morning’s “woof” page and consider its significance. I for one cannot muster even a derisive chuckle over Target’s epic fail. It’s just too serious a matter. I mean, this is your brand on the line — or off the line, as the case may be.
As a matter of fair play, I should duly report that as I was completing my previous post (8:22 AM PDT), the Target.com site came back online.
With revenue on the line, Target will get this right. Consumers are generally forgiving, so I wouldn’t anticipate any long-term hit to the brand.
I can’t think of any great company that hasn’t botched a launch or misjudged the market. Even Apple. Ever hear of the Apple Newton?
Target’s web site has been down and now up today. I don’t understand why a company with the depth of resources, and funding to buy what they do not have, could have bungled this so badly.
As Doron said so well a glitch is a fancy IT term for failure. Companies of my size have glitches and customers will understand. Companies like Target; it is not understandable.
The challenge with implementing real time IT applications has always been “you don’t know what you don’t know” so there will always be surprises. This has become even more critical when you flip the switch on a customer facing application. The additional challenge in this situation seems to be the coordination of five different vendors and the complete transfer of the service at once.
For the role out of traditional retail updates I have always advocated a two phase approach. Basically, I recommend changing all the internal support mechanisms such as the data model, new hardware, and new communications channels separately from the functionality and user interface changes. Let the IT department work out their internal problems without impacting the user training and operations aspects of the business. This often means writing some throw away code that bridges between old code and the new infrastructure but it lets the IT department make sure the backend can support the new environment. Once the old applications are running on the new platform it is time to convert the user interface and controller aspect of the application. If problems occur, the backend can be restored from the latest backup and the old frontend put back in service.
In this particular situation the magnitude of change seems so great that a phased approach was not necessarily possible. In this case, the goal should have been to ensure all the functionality was in place and that the only concern remaining would be “can the system scale to the necessary volume?” The standard quality assurance testing should have been followed by a limited beta that enlisted all the Target employees to make purchases over a beta of the new site. Log files would be used to monitor beta operation to ensure all the parts of the new application were tested. Finally, with only scaling being the outstanding issue, the cutover to regular customers would be made.
The good news is that Target made the cutover now, in two months they would have impacted holiday traffic. If the site can provide the new functionality and prove more user friendly in the long run, Target will benefit. Customers are quick to forget. If users of an IT system are forced to live with chronic problems, there is much more frustration. On the other hand, if the issues are quickly resolved and the new features are appealing to customers they will soon forget the start-up problems. Besides, a store credit goes a long way to make up for inconvenience.
Last week someone took the words out of Warren’s mouth, and today Doron and Richard took the words out of mine: sure glitches are to be expected…when you’re a 20 year old newlywed, or a group of parishioners volunteer to run a bake sale. I would have hoped Target would have put more effort into this.
Expect more, pay less? No; expect more but pay less, get less.
Well it is now 11:30 pm East Coast time and the site appears to be mostly working. It was down when I first tried and them came up 10 minutes later with everything functioning except the “daily ad” link.
The comments above are accurate:
1. “Glitches” are certainly failures, and
2. The customer will not hold this against Target.
It is, however, a significant embarrassment to the IT team and should be a cause for concern for senior leadership.
Overall I like the design of the new site and think the situation will be forgotten by tomorrow — as long as it continues to function!