NFV’s Success Depends on Organizations Adopting a Fail-Fast Culture
Part of the promise of NFV is that it will make enterprises and Communications Service Providers (CSPs) notably more agile. In the case of CSPs, it is often said that NFV will make these organizations as agile as Over The Top (OTT) players such as Google. I have no doubt that implementing NFV will make organizations more agile at performing the same tasks they currently are doing. For example, NFV will enable an organization to dynamically add the capacity it needs to meet its service level agreements for an existing service. The bigger question is will NFV enable organizations to roll out new services notably faster? The answer to that is yes it can, but only if the organizations adopt a fail-fast culture.
The concept of fail fast refers to spending less time planning new products and services and focusing more on getting those products and services into production quickly. It also refers to acknowledging key learnings from whatever failures do occur and leveraging those learnings to quickly move forward with new products and services. The concept of fail fast is the subject of numerous articles and books. Projects such as Google Glass are often held up as the poster child of what it means to fail fast.
I used to work for NYNEX. NYNEX was a US-based CSP that merged with Bell Atlantic which later morphed into Verizon. At NYNEX I was the engineering manager for new high-speed data services and the process for rolling out those services was extremely lengthy. One of the reasons it was lengthy is because the marketing organization wanted to do as many studies and interviews as possible to ensure that their projections for how much the new services would be used weren’t wildly off the mark. The equipment I was deploying was purpose built for each service. Hence, we tended to evaluate and then re-evaluate that equipment and the enabling technologies because if the new service failed for whatever reason we wasted multiple millions of dollars and there is no doubt that that would have had a negative impact on our careers. In addition, the people who ran NYNEX’s central offices were in no hurry to implement new classes of equipment. These people were evaluated on their ability to meet certain operational metrics and introducing a new class of equipment made that very challenging.
The good news about the current environment is that new technologies and architectures, such as NFV, totally change how an enterprise or a CSP can deploy new services. For example, a CSP can have generic servers in its central offices. It can then quickly download software onto those servers that enable new services and the CSP can pay for that software on a pay as you go, pay as you grow basis. As a result, if the service bombs in the market, little or no money is wasted.
There is no doubt in my mind that because of adopting NFV that CSPs will soon be able to roll out new services without taking the multi-year approach described above. How much faster CSPs will roll out new services is still a big unknown. My description of merely downloading software and being able to offer a new service is clearly very simplistic. CSPs still need to do testing and integration and they will still want to tie into key systems such as their billing systems. There is no doubt that these factors will slow down the implementation of new services, but how about the corporate culture? Will the people in marketing, engineering and operations take a fail fast approach and push to get new services out quickly or will they hold back, wondering how much their year-end evaluation will be impacted if those services are less than 100% perfect the day they are launched?
A fail fast culture is relevant to more than just CSPs. I used to work for is Digital Equipment Corporation (DEC) where I was in charge of transmission, switching and routing for DEC’s internal data network. Shortly after I took on that role I had to make a presentation to DEC’s IT leadership team. In that presentation I said that I wanted my group to succeed on 9 out of 10 new projects. I made that comment in an attempt to set a direction for risk taking within my group. The management team thought I was joking and they laughed. I quickly realized that within my organization acknowledging failure was not an option. Every project, no matter the outcome, had to be positioned as a success.
I think it’s fantastic that CSPs and enterprise organizations are spending so much time and resources evaluating NFV and conducting POCs. However, I believe that Peter Drucker was 100% correct when he said that culture eats strategy for breakfast. To me that means that if CSPs and enterprise organizations don’t also spend time and resources changing the organizational culture, that they will not experience the full benefits of NFV.