Labs Strategy: Why Embracing "Failure" Is A Great Idea But A Horrible Word

Oracle posted a very interesting short thought-piece Wednesday (June 19) about the different ways retail chains do—and should—handle lab strategy. The choice of strategies reveals an awful lot about the attitudes of senior management. One key point is that management must be willing to accept—and learn from—failures. But if the CEO even thinks of the ultra-valuable data that comes from learning what shoppers will not accept as "a failure," the chain has even bigger problems than it thinks.

By looking at the different choices made by Walmart, Target, Home Depot, Nordstrom, Staples, Tesco, Wet Seal and Lowe's, Oracle categorizes three IT lab approaches. But how a lab is corporately structured will make little difference if senior management isn't willing to first learn (and to pay a lot for those lessons) and to be open to a future that they may not like. The job of a chain is to adapt to the reality in its market. The job of a dying chain is to cling to its current tactics if the future doesn't look like what it wants it to look like.

Labs are often pure internal mechanisms, but they are more often the result of a tech acquisition (Walmart [NYSE:WMT] buying Kosmix, Home Depot [NYSE:HD] buying BlackLocus and later Red Beacon).

As Oracle's David Dorf (senior director of technology strategy) penned: "In most cases, these labs stem from acquiring a start-up, and not wanting to crush the start-up spirit, the retailer keeps the company separate." Integrating the techs into existing groups is indeed a bad way to crush their start-up spirit, especially when chains like Walmart and Target (NYSE:TGT) have so many more effective ways to crush their spirits.

Although Dorf has earlier also talked up the labs at Nordstrom (NYSE:JWN) and Staples (NASDAQ:SPLS), Wednesday's piece focused more on the different ways to strategically use these labs. The piece has an interesting graphic that looks at just the acquired components of Walmart Labs and how it's integrated no fewer than seven companies (Kosmix, OneRiot, Grabble, SmallSociety, TastyLabs, OneOps and Inkiru) into its lab.

Dorf argues that the acquisition approach is fine, provided "there's no existing culture of innovation, so buying the start-up mentality can form a basis for building a lab." That's an interesting point, but one can easily argue the opposite, that when building through acquisition, it is critical that the chain have a very strong and well-defined existing culture of innovation. Otherwise, chaos and duplicate (wasted) effort is almost guaranteed.

A single acquisition can work, if the company is prepared to embrace that culture. But when the acquired growth includes several companies, there needs to be a strong top-down direction and it needs to be well articulated. One of the challenges is that retail IT, in general, struggles with effective communication. And nowhere is that effective and consistent (and detailed) communication more essential than when trying to integrate multiple teams with entirely different cultures and approaches.

Add to that the psychological factor of the typical acquisition. Rank-and-file workers (which is where so much of the tech magic happens and truly needs to happen) fear (or, in some cases, pray) that their old management will be overruled and made irrelevant. Without clear direction that is well-communicated, workers will be hesitant to follow their old leaders and will be unclear about their new ones. In an idea lab, passion and enthusiasm are essential. That's something that demoralized acquired workers rarely have. Oh, they'll do their jobs, but with one eye on e-mails from headhunters.

A different approach, Dorf wrote, is an organic approach, along the lines of what Tesco and Wet Seal have done.A different approach, Dorf wrote, is an organic approach, along the lines of what Tesco and Wet Seal have done. But he then taunts chains: "This approach requires strong leadership, vision, and a willingness to fail so it's not for every company."

Strong leadership and vision is necessary for any of these strategies, especially the acquisition route, so I'm not seeing that as a differentiator.

As for having "a willingness to fail," that is also certainly something that is needed for all of these approaches (so, again, no differentiator). But "willingness to fail" means several things. It usually means a willingness to see a bad outcome as not "failure" at all, but a useful piece of information. For example: "We need to know what will work and what won't in (fill in the blank – mobile, social, geolocation, etc.). The fact that this approach did not work is great, as it told us what we shouldn't deploy. Thanks, team. This is extremely valuable data."

This presumes that everything has been properly analyzed so the proper conclusions are drawn. Did it not work because we're too early? Was the deployment flawed but the idea valid? Did it not resonate because of the geography tested or the demographic that we used?

Part of the challenge with any tech trial is that the trial is based on a lengthy series of assumptions (guesses) that the team made. If it doesn't work, it's almost impossible to truly know why. Without knowing why a trial delivered or didn’t, it's very difficult to make the right extrapolations from the data. One thing is certain, though: A trial that didn't deliver invariably delivers much more valuable data than one that succeeded.

The key point is that if management is even using the word "failed," it's a pretty strong heads-up that they have no "willingness to fail." A trial is a research tool, a learning experience. To equate a negative outcome with having failed misses the whole point of a trial.

(In the media business, useful feedback is a rare and beautiful thing. But nice and happy comments [Great work! Keep it up] are much less useful than specific negative thoughts, assuming our goal is to constantly improve what we deliver. That's how trials should work.)

The third category Dorf wrote about was "partner collaboration" and he cited Lowe's (NYSE:LOW) quite properly as a good retail example of that approach. Dorf's words here are precisely on-target: "The danger retailers face is losing focus on their core competency—retailing. Running a start-up within a large company can be costly, reliant on key individuals, and sometimes a distraction to the core business. An alternative approach is to partner with technology companies so as to share some of the burden. Lowes, for example, invites technology partners to present innovative ideas then chooses a few projects for collaboration. This can be an excellent way to stay on the leading edge of innovation without some of the mentioned downsides."

The only nuance missing here is that most chains have two conflicting pressures that it's up to the COO and CEO (and, ideally, the CFO) to resolve. While all of this change and exploration is going on, store and online management must focus on the day-to-day business of running a profitable business.

The CEO message to LOB managers should be: "People, we're going to be exploring and toying with a lot of things and some of those trials will be happening in your sites and stores. I absolutely need you to shout and scream if any of this in any meaningful way starts to undermine your goal of running that profitable business. I'm going to be trying a lot of things and I'm relying on you to tell me if it hampers you in a dangerous way."

But the opposite is also true. For a chain to grow and to be effective 2-3 years down the road, sometimes day-to-day operations and profits must be sacrificed for the future, for the greater good of the chain. That's where the willingness to fail (I still hate that word) becomes so critical. How can an executive know whether the current non-traditional experiment is wasting time or a valuable research tool?

The way it's supposed to work is the CEO gets lots of feedback from experts who report to the CEO. But there are historical biases: CIOs embrace the future and generally love to experiment while store management is scared of those efforts and sees them (sometimes correctly) as distractions that waste time and money.

Bottom line: It's not as much a culture of research (willingness to fail) that is essential as a COO/CEO who are confident enough in their decisions to make those moves—and to not necessarily see a negative outcome as proof that they were wrong.