“Sudheesh, I will level with you. I like the Nutanix technology, I like how you guys have conceived, designed, and architected the product and I even like you! I know I really could use the increased performance, scalability, and cost savings in my datacenter.”
Uh oh…I knew from past experience that anytime a prospective customer starts a conversation singing our praises, there would be an inevitable “but” making an appearance real soon. I also knew that I wouldn’t like it. I waited. The IT Director at a medium-sized financial institution leaned back in his plush leather chair, fixed his sharp eyes on me, paused for a second, and continued.
“But unfortunately we can’t buy it because I am sure that you guys can’t provide us the kind of support my current big vendor is providing.”
You couldn’t blame an emerging technology company sales person expecting to wake up from a recurring nightmare in a cold sweat at this point. The question of support is a common and often very effective objection that customers have against an early stage company and its products.
Taken at face value, this is an insurmountable objection. The company in question, such as my employer Nutanix, probably is a private company, probably has not been in business for more than a few years, and certainly won’t have the kind of resources that a large company like EMC or HP can deploy all over the world.
But let’s take a deeper look at the facts behind this objection. After all, support is not just about the size of the company. It’s about hardware, software and, most importantly, about people.
Fact 1: Hardware will fail and software will have bugs
If any sales person tries to make a case that you don’t need world-class support because the product is designed so well that you will not need it, stop immediately and call for security to escort that person out of your office.
If anything can go wrong, assume that it will. It doesn’t matter if the product is from a 1-year-old company or a 100-year-old company. “Nothing will go wrong” is not a valid answer to the question of a company’s support capabilities. Yes, over a period of time bugs and problems will level off. But according to Mark Leslie (a Nutanix board member and former CEO of VERITAS), such a leveling off could take up to 10 years. If you aren’t planning to refresh your technology for 10 years or more then this has the potential to work. But that isn’t a feasible option.
As far as I am concerned, this is a playing field leveler for both startups and big companies. Any customer would, rightly so, want to make post-sales support an important criterion while making decisions on your next generation data center infrastructure.
Fact 2: Software support is the key
Today’s datacenter is full of very complex software running on relatively stable and powerful, commoditized hardware—but software is where the real support complexity lies. For example, most storage systems will have anywhere between 1 million to 10 million lines of codes.
Most companies consider serviceability to be luxury that startups can’t afford. At Nutanix we consider it essential for laying foundation to a lasting company. We have taken a two-pronged approach towards support.
First, we designed the system to be intelligent. One of our smartest engineers in the company, who was developing hardcore storage systems at IBM Research before joining Nutanix, is responsible for the design and development of our serviceability module. Our system has a dedicated subsystem to monitor hardware and other key software subsystems within the cluster. It will proactively alert our customer support team if it senses an anomaly. This system is designed to be capable of machine learning through intelligent analytics: that is, the more we use it, the more intelligent it will become.
If the serviceability subsystem senses a problem that requires active troubleshooting by an engineer, it will package up all the necessary activity logs and test results without you having to do it manually. (Of course, this feature is not always on and can be disabled at your command.)
Second, we have invested in people. Chances are you don’t call customer support when things are humming along nicely. You call when you are sure that something has hit the fan. What you need is your problem solved; what you don’t need is to talk to someone who is either acting as a glorified stenographer or insulting your intelligence. This is where a well-run early stage company can truly differentiate themselves from the large players.
At Nutanix, your level-1 and level-2 support are handled by a very smart group of System Reliability Engineers (SREs) based out of our San Jose, LA, and London offices. These are guys and gals with advanced degrees from top schools and are passionate about customer support. We have people from VMware, Accenture, and other organizations of their caliber. I am a strong believer that one of the reasons Nutanix has seen so much repeat business is because of the high quality of interaction our support center provides.
Now if you have a problem that needs to be escalated to level 3, you are probably going to interact with the engineer who actually wrote the code. We all know that developers like to write code, but don’t really enjoy taking customer support calls. At Nutanix we have implemented a system where the on-call engineers will receive a substantial bonus during their rotation. The result? You will actually be talking to developers who are eager to help you and looking forward to a happy and fast resolution to the problems.
Can you say that about your support contract from the large vendor?
Fact 3: Great Hardware support is about simplicity
Good clustering technologies are built on a simple theory. It’s called “fail hard, fail fast”. Simply put, if you sense trouble with any component, don’t try to hang around with it too long but simply sever the relationship with it and move on to rebuild the cluster with what’s left. (BTW: This method is also highly recommended if you are single and not looking for serious relationships and stuck with a “bad component”.) With this model, if our cluster manager software senses trouble with any component, it will shut it down, send alerts and continue humming along with what’s left. That is the beauty of an N-way scalable asymmetrical clustering architecture.
Nutanix is designed to have no single point of failure and to keep most of the field-replaceable units accessible without taking the system down. Moreover, there is nothing in the hardware that you probably haven’t seen before. This is important because the cluster is a dynamic entity with ability to shrink and expand when a node failure/replacement procedure is done. A number of our customers buy our spares kit and do these procedures themselves.
However, Nutanix has also partnered with a third party to provide onsite parts replacement anywhere in the world. We have established SLAs with this third party to exceed the SLA expectations we are setting with our customers. Nutanix won’t sell support to customers if we can’t establish a local warehouse to meet our committed parts replacement SLAs.
Fact 4: Converged architecture helps us!
Customers hate it when vendors start fighting with each other in the middle of troubleshooting a problem. How many of you have experienced a situation where you were told “it’s not my problem, call the other vendor”? No matter how simplified a stack looks, there are so many drivers and configurations in today’s systems that it is all too easy to cause a problem that is difficult to troubleshoot.
The genius of a truly converged architecture is that anything below your main datacenter network is Nutanix’s problem. Your compute resources, storage resources, VMware issues, vCenter issues, broker configuration issues for VDI. You name it, our customers call us about them. This is why our support team is not just good at Nutanix technologies, but also at the entire virtualization ecosystem software and hardware that constitutes it. Customers always have one throat to choke: ours!
Fact 5: You still won’t trust everything I am saying
Let’s face it, I am a sales guy who won’t make my mortgage payments if I don’t make the sale. You know it and I know it. So you aren’t going to trust everything I wrote here. You still want to know if we are going to lock you down with a proprietary architecture. This, again, is a fair question.
All of the intelligence Nutanix provides is contained within a VM running inside your VMware hypervisor. Moving into Nutanix is a simple matter of Storage vMotion or cold migration. Moving out of Nutanix is a similar simple migration.
Let’s reflect on all of the aforementioned points.
- The hardware, enterprise x86, is something you know very well.
- The base OS is VMware that you have in your datacenter, and Nutanix doesn’t modify it or write any crazy drivers for it.
- You can monitor, maintain, and mange the cluster from VMware vCenter.
- Hardware parts replacement is pretty straightforward and includes call-home with global logistics support.
- You will be working with an intelligent system and very intelligent support people and engineers who will gladly support your entire virtualization ecosystem.
- You will have access to the developers and executives in the company if you want to escalate anything.
Tell me again, where is the risk?
When you are selling for an emerging technology company, a lot of things go against you, such as brand recognition, marketing budget, size of installed base and lack of executive sponsorship in most accounts . But, at least when it comes to Nutanix, customer support is not one of them.
Once a potential customer asked our brilliant support lead if we could provide support like their big vendor can. He answered “No, but why would you settle for that when you can have the kind of support only a small vendor can provide?”
We care about every customer, big or small, because we have a lot more to lose than any big vendors do. In this highly social-networked and connected world, a customer with a bad experience in Ohio can mess up a deal for us in Australia. At Nutanix we realized early on that the bar is much higher for a new player. Fortunately that is exactly where we like the bar to be.
PS: The aforementioned customer who had serious doubts about our ability to support them is now a very happy repeat customer for us and done a lot of reference calls on our customer support.