New CIO cleans up outsourced IT at Clorox

The Clorox company brought in a new CIO to re-establish control of the company’s tech direction, bring IT up to date and generate business value.

The $5.5 billion Clorox company brought in Ralph Loura as CIO last April because, among other things, it realized it wasn’t getting what it wanted out of an extensive HP outsourcing deal. In quick order, Loura and his reformed leadership team fixed some outstanding tech issues and re-established internal control of the company’s tech direction. Loura talked about what he found when he joined, what he has accomplished so far and what he’s planning next with Network World Editor in Chief John Dix.

By way of background, Clorox is a 98-year-old company based in Oakland, Calif., that has 30-plus global brands. These brands include the namesake bleach and everything from Kingsford Charcoal to KC Masterpiece Barbeque Sauce, Glad bags, Hidden Valley Ranch Salad Dressing and Burt’s Bees, a natural personal care product line. The company employs some 8,300 people, about half of which are plant-based, and has 120 people in IT. There are another 300 to 400 IT people working on the Clorox account at HP.

NW: You were brought in as CIO of Clorox to set a new direction in IT. Tell us about that.

R: Clorox had essentially viewed IT as a cost center, a cost of doing business, and the last couple of CIOs had come from a finance or audit background. They were multiple years into an outsourcing relationship with HP and were a bit frustrated that the productivity improvements and efficiency gains they expected weren’t materializing. What they came to realize is outsourcing is not a turnkey operation, that the business has to lead some of the value generation, so they went out looking for someone who had prior experience as a CIO leading business transformation and generating business value through the deployment of technology.

IN PICTURES: Where the IT jobs are: 10 American cities

NW: What did you find when you got there?

R: What I expected. An organization that had been under-invested in, both in terms of people, process and technology. For example, our desktop standard was Windows 2000. Most organizations had at least made the move to XP, if not to Vista. And some of the underlying systems, our ERP environment, our CRM environment, had become dated. Another example was e-mail messaging. We were a Lotus Notes shop and running a version of Notes that was several full releases behind.

Everything was in need of basic upgrades and maintenance to bring the environment from a dated state to a more current and more agile state. There had been this kind of, “If it isn’t broke, don’t fix it” view, which is contract behavior, right? You do the least amount necessary to maintain the SLA, with no idea that certain upgrades create value by unlocking productivity through the features in the solution.

NW: So you come in and find a bunch of old stuff, a lot of projects screaming for attention, how do you figure out what to do first?

R: Well, some of this is opportunistic. So clearly in the first 90 days you want to make an impact. And an impact doesn’t mean starting a portfolio of multimillion dollar projects that will take three years to pay out. You need to move the needle. You have to triage what can be done quickly, what can be done with the highest impact and lowest risk. So one of the first things we did was start a project to off-load Windows 2000. And again, because we wanted to pursue both speed and low-risk, we actually moved to Vista. Some people questioned why we didn’t go directly to Windows 7, but we looked at Vista as a tactical upgrade to buy us some time to make the real strategic transformation that’s now in the pipeline.

Our intent isn’t to upgrade our desktop image to Windows 7. Our intent with the Windows 7 lever is to fundamentally change the way we deliver and use computing services at Clorox. Leveraging VDI thin client technology, extending self-help, having much more of a self-service model around app deployment, and so on. All those things are enabled by Windows 7, but we didn’t have time to engineer and deploy that level of change, so we did the Vista migration first.

With the advent of more cloud-based offerings in the marketplace, we were also able to quickly migrate e-mail to a Microsoft hosted Exchange environment. Instead of doing a 12 to 18 month internal migration project, building exchange servers and migrating the user base and so on, we were able to do several months of testing followed by literally four weeks of cut-over to move the entire company to the hosted Exchange environment, what Microsoft is now calling Office 365.

NW: Any other quick hits?

R: We embarked on a point release upgrade of our core ERP environment, and we did some work in HR and some other things, but I’m a firm believer that IT doesn’t happen in a vacuum, that it has to be base-lined to the business strategy. So we began a listening tour, traveling throughout the company listening to what people had to say about what pleased or frustrated them about the IT environment, and then attempted to create a prioritized set of tasks we could tackle.

What became apparent rather quickly was we had to categorize the size and volume and type of work into multiple buckets or we would drown all the little things that were annoying but didn’t have a topside ROI that was compelling. So we created four categories of projects: business transformation, project road map, project short cycle and break/fix/enhance. And we began slotting projects based on size or impact, time to complete, or cost into one of those buckets.

And there were some things that were business transformational in nature. We identified, for instance, a compelling need to consolidate and support a more consistent set of ERP platforms internationally because we have substantial growth in that business but it is being supported by 15 different ERP systems across 18 countries. Rolling up a quarter consistently was difficult because there were too many different systems in play. And the underlying support costs to maintain changes and support for multiple instances tend to pile up. So that categorized as a business transformation effort.

There are other things that are more incremental change and improvement, like enhancing our underlying network infrastructure, and then there are these short-cycle projects that are sort of the hearts and minds campaign, if you will.

We ended up with several hundred, pared that down to about 120, then sequenced that down to about 90 items that we began working through a service life-cycle management process. And what’s really nice is we’ve gone from an environment where virtually every project ended up over budget, over time and missing some core functionality, to an environment where, for the most part, we’re within plus or minus 5% on the smaller items of being on time, on spec and on budget.

We’re pretty proud of that. We think we can do better, but we’ve essentially got credibility that when we say we’re going to do something at a certain cost and within a certain time, the business believes us and are willing to allocate resources to support that effort because they know they are going to get a tangible value at the end of it.

NW: How are you changing your relationship with HP?

R: I think we relied on HP for things we shouldn’t have relied on them for. You can take HP out and put in IBM or any other large scale outsource provider; it’s not the specific provider problem, it’s the model. We expected too much out of a single outsource provider. So we elected to restructure the function and bring three key roles in house.

One role we always had but it wasn’t structured for success was the role of business systems manager (BSM). We’ve rebranded it as client manager and refocused it to be more about the client and less about the system. Before, there was a strong focus on systems, say, the ERP system. Now their job is to own the client needs, make sure they are met, and we’ll support them using whatever system is appropriate.

Another role we had was the role of architect, but the architects did anything but the stuff architects need to do in a large-scale IT organization. They spent most of their time at a tactical level, going from project to project, building bridges between people and systems. We hadn’t effectively planned what our overall environment looked like so we had a lot of problems creating these bridges and the architects didn’t have any time left to go look at, for example, cloud HR services versus on-premise options.

So we created a role called the solution designer and they essentially do what the architects were doing before; they’re tactically engaged on an ongoing basis with projects and help client crews leverage the feature and functionality of various applications. That frees up the architect to do that longer-term road map and vision work.

NW: Did you have to bring in some new folks as well?

R: Yes. We’re proud of the fact that we had very few people leave the organization — less than five — and we opened just over 20 new roles, so net new we added 15 to 16 jobs. In most cases we had people with the talent, the experience to play the roles we were looking for, they just hadn’t had the clarity and focus to be able to go pursue that activity. So by bringing in these additional folks and realigning some job responsibilities, we were able to generate value in a more meaningful way.

NW: Otherwise are you happy with your outsourced situation?

R: You know, like anyone, we have challenges with our provider. In some places they are aligned well and delivering well. For instance, we’ve dramatically improved customer satisfaction scores across the help desk by us providing a bit more direction and guidance, more clarity in decision making, providing them a framework to better respond to users, and I think that is fairly solid.

There are other areas where we continue to struggle a little with delivery or skill set in the outsourcer. They are, I think, chaffing a little over us coming in and being a little more directed in some areas, taking more of the architectural strategy reins, if you will, and asking them to step up in a different way than they had before. We’re still working through some of that.

NW: OK. Let’s turn back and dig a little deeper on those early projects you talked about. When will you start with Windows 7, and can you explain more about what the goal is?

R: We just officially finished the Vista migration in November, and in December started planning the Windows 7 rollout. It was essentially a matter of waiting for the team to complete one activity before they could be focused on the next.

With the traditional model of deploying desktop computing all the systems start out relatively the same, they have the golden image. But literally the first day users log in you start to see deviations. You get a different set of apps for each user, they’ve changed their registry settings or what have you based on what’s connected, and all of a sudden you’ve got 5,000 unique devices to support instead of 5,000 nearly identical devices to support.

So the vision for Windows 7 is to load software in layers. The application layer of abstraction Windows 7 allows is much cleaner than any prior version of Windows. So I can deploy a stable OS base that I can virtually guarantee will always be stable as long as it came out of the lab that way. I’m not going to pollute the core image with a bunch of DLLs every time I load new software because I can abstract the software I’m going to load into a kind of virtual memory bubble.

That provides stability and dramatically lowers costs to support the desktop environment, and that’s the core vector we’re going to try to accomplish with the Windows 7 rollout.

In addition, because of how that works, we believe we can eliminate a lot of things that people call the help desk for today by getting users to click on Web links to download software or to verify configuration or what not.

NW: Any plans for VDI?

R: We don’t currently plan to deploy highly virtual or ultra-thin clients, although we are really intrigued by them. So we’ve got some work in the lab going on, as part of the exploration we’ve got to do before we finalize our Windows 7 rollout strategy. There are some cases where we think there might be a good fit for ultra-light, ultra-thin, VDI environments. It might be a good model for our plant-based workers, for example, because they could log in from a kiosk in the plant between shifts or go home and connect from essentially any PC and run applications in that corporate cloud.

NW: How many desktops are you talking about anyway?

R: We’ve got about 4,800 inside the company.

NW: Speaking of cloud, what did you learn about migrating e-mail to that hosted Exchange offering?

R: The most surprising thing was we presumed we would have to do minimal training on Outlook because it is the most popular e-mail client on the planet. What we found was a surprising number of people needed some basic Outlook training because they had either been at Clorox for a decade or more or hadn’t worked at other companies that used it and certainly weren’t using Outlook at home. We actually had to delay our original deployment plan by about three weeks while we went and developed a richer set of training materials.

NW: How are you liking the hosted version?

R: So far we have been quite pleased with the service, the performance. Users rave about some of the features and functions we unlocked, like being able to sync from pretty much any smartphone.

NW: What are your bigger expectations for cloud computing?

R: First of all I’m going to try to avoid using the word “cloud” because it doesn’t really mean anything to anyone anymore. It has morphed in so many different ways.

But we believe in virtualizing essentially anything we can. Our network that connects all domestic sites, for instance, is an MPLS-based network, so you can argue we’ve virtualized our network. And in our Windows environment, we’re inching up toward about 70% of our Windows images being virtualized.

At the infrastructure level we’re highly virtualized today in a “private cloud,” one we built. I think over time the natural evolution will be toward buying instances in the cloud in a multitenant, virtualized, platform-as-a-service kind of model. That makes sense to me for certain workloads, such as development, QA, things that don’t have high risk or require high performance or low latency.

But I think the real win has been and continues to be those tightly integrated vertical apps like Salesforce.com that have a loose coupling to the rest of your enterprise. I’ve deployed Salesforce at two prior companies and we looked at Salesforce here for some international environments, but they don’t currently support some of the functionality we need in the consumer goods space. But we are looking at a couple of different players who are starting to build those solutions in the Salesforce cloud to see if there’s value in pursuing that model.

In other cases it just doesn’t make sense. Doing data analytics in the cloud, for example, seems very odd to me because of the large volumes of data that would have to be extracted on a daily basis and uploaded to a cloud.