By Jahidul Khandaker, CIO, Western Digital
When I first started in IT, my boss told me never to fall in love with any technology.
His advice has served me well, particularly over the past four years while working on the integration of three $7 billion-companies, SanDisk, Hitachi Storage Company and Western Digital. When the three organizations came together, we each had our own full-scale infrastructure, systems and processes. Since then, we have been trying to consolidate our infrastructure and 3,000 applications as well as drive a digital transformation of some key processes, all at the same time—jobs that would have been much harder to complete if I had let myself become attached to particular solutions.
Start with questions
We asked three questions about each application in our three portfolios. First, is this a legacy technology of one of the three companies that can serve our entire, new $18 billion company? If we decided that, yes, it was the right one to adopt, we chose it and scaled it to serve the whole enterprise. For example, two of the three legacy silos had Workday, so we made one of the Workday iterations primary, and then scaled it out for the entire company.
Then we asked ourselves a second question: If we had a technology that was good but that we couldn’t necessarily scale to stretch across the entire company, could we reimplement it in a way that would make it possible to scale? That took care of some other application.
Finally, if none of the technology that we had could do the job, and we had to buy something new, what should we buy? For example, for our ERP backbone, we had two instances of SAP on premises and one instance of Oracle EBS, but we decided to go with one instance of Oracle Cloud. We are moving the entire company onto that now, and we have been making very good progress towards that consolidation.
Ultimately, we expect to end up with around half the number of applications we had when we started. We are probably about two years away from achieving that goal, but we have already eliminated hundreds of applications from our footprint.
Listen to the business
At the same time, staying emotionally uninvolved with the outcome doesn’t mean you should discourage a serious debate about the choices that need to be made. Here, the business units need to play a very critical role—their expertise and resources are key to the success of any transformation. In a large transformation, you almost need their full-time participation.
Their participation is also important because the needs of the business units can change quite quickly. The push toward more remote work this year, for instance, may be pushing us toward a more purely digital environment faster than would have otherwise been the case.
When we realized how important processes were going to play at a company level, we also established 14 end-to-end processes, and for each of those processes, we identified a global process owner—a single owner for each end-to-end process, even if it crossed functional boundaries or legacy boundaries. By giving one person authority to look after and optimize an entire process, we were able to speed up integration substantially for some of our most important workstreams.
The pitfalls of automation
We’ve applied each of these principles to our large-scale automation program, which is key to digital transformation success. With numerous islands of automation, we need a smart enterprise approach – becoming efficient at the wrong processes would only slow down our transformation. We are still very early on that journey, but with our streamlined application footprint, the foundation is in place to unlock enormous value.
Last year, we set a 12-month goal of reducing 100,000 hours of labor through automation, and we have already hit that target. We are a large company—we have about 65,000 employees worldwide—so automating even a small fraction of our transaction processing, our business processes, or any activity that we repeat yields significant savings.
This is just the beginning. We will continue to work with the business units to find more opportunities to implement, and I believe there will be many more. It’s an extremely valuable exercise: beyond saving on labor costs, automation improves accuracy and reduces cycle time. It also allows more people to focus on high value transformative work rather than repetitive transactional work, and that itself brings more value.
Automation is very tricky, however. We have to be very careful, making sure that it’s secure and foolproof, because when you are automated, you are doing thousands of transactions really quickly. If you are doing something important like order creation, getting it wrong is not a good thing. We need to make sure that the governance is there, error handling is there, and the integrations are there.
End of a love affair
Of course, extending capabilities such as automation means letting other things go, as you would in an ordinary integration. It means not falling in love with a particular system or legacy process. Accepting that there will be continuous change has enabled me to work on this integration and automation project objectively without having to worry about whether we were replacing a Western Digital process or a SanDisk process. The technology keeps on changing, but two decades later, my old boss’s advice is still serving me well.