Last month, I suggested that VMware could help itself and its customers by opening up. How about if it transformed completely? Could VMware survive if vSphere were free and it only charged for support? What if VMware embraced the idea that hypervisors are a commodity and focused on revenue from management and automation products? What if vSphere didn’t have a lot of different editions?

VMware needs to change. We’re seeing large customers reducing their vSphere footprint. Those same customers are moving workloads to public cloud. Many are also adopting alternative hypervisors that have lower up-front costs. One of the challenges that VMware needs to address is that cloud services are billed monthly, rather than once every three years. Cloud services also bill for consumption, which is usually related to how well the business is doing. A successful business is happier to pay a bigger bill for a product that helped it achieve that success. Customers like a monthly bill that is directly tied to business value, rather than a less frequent ELA bill that makes no sense.

If you only charge for support, then your support must be awesome. When I first started working with VMware products, ten years ago, the support was amazing. Sometimes, the first person who answered the phone call could resolve the problem. If the problem was tricky, it might only take two calls before you were talking to the person who developed the feature you were fixing. Support was outstanding and worth every penny. This level of customer support is very hard to scale as you get a massive increase in installed base. Those developers still need to write code, so they cannot be handling every support call that comes in. Over the years, VMware had to massively expand its support team to handle the enormous growth in customer numbers. This was a bit of a rocky road at times, but now the VMware Global Support Services (GSS) organization is huge and very systematic. GSS engineers can resolve a lot of technical issues without needing to escalate to developers. VMware’s support is definitely good enough to justify a serious price tag.

Maybe the change doesn’t need to be a shift to only selling support. A lot of VMware’s vSphere revenue is already from SnS renewals. SnS is basically a third of the purchase price of the product. You pay each year, and as long as your payments are up to date, you get new versions when they are released. So why not remove the up-front vSphere purchase cost and just have an annual SnS fee? Better yet, allow an option for monthly billing of SnS. Version upgrades could be tied to having current SnS paid up. Some elasticity in the SnS would be great, too. Allow customers to deploy ESXi servers for a few months and only pay SnS for those few months. Tie the fee that customers pay to the value delivered.

Another option is essentially a bundle. Make vSphere free with the purchase of management products like vRealize Automation. This is the model that Microsoft uses: Hyper-V and basic management are effectively free. For comprehensive management, Microsoft sells System Center Virtual Machine Manager (SCVMM), which is far from free. To get the equivalent for vSphere, VMware could have a free but minimally featured vCenter and ESXi combination. The free level could include both vMotion and vSphere HA. It would essentially replace all the vSphere stock codes below Enterprise with a free tier. Naturally, support would be a paid option. Enterprise customers would still buy Enterprise Plus. Most commercial customers would also buy SnS, so VMware would keep making revenue. But Microsoft would have far less opportunity to talk about the VMware tax. Customers would have more opportunity to use vSphere without huge costs.

A smaller-scale change might entail significant price reductions on vSphere, relying on the effects of the Jevons paradox to make more money. The Jevons paradox says that increasing the efficiency of resource usage can lead to greater demand for that resource, and therefore more profit. If a price drop leads to significantly greater usage, then VMware could increase its profit by reducing its price. Alternatively, it could slow reduction in revenue by reducing the value of switching from vSphere to cloud services or another hypervisor.

It is time for VMware to do something different. I’m not convinced that VMware is going to win against containers and public cloud services. Changing the purchasing model for vSphere could be overdue. It is possible for both VMware and its customers to be better off due to a new way of buying a core infrastructure component.