Estd 4.22.2023
.png)
.png)
Software related to IT infrastructure and Operations Management has been historically priced based on hardware or asset size under management, be it the number of processors, the size of the storage or the amount of data transferred. In this model, a customer would buy a license for a particular size and then try to stay within the limit imposed by the license. The vendor would also struggle to ensure that the customer did not use anything more than what the license permitted.
Public cloud providers like AWS and Azure heralded a new approach with architectures that allowed their customers to provision hardware and software in a granular and elastic manner. This model has since been taken up by almost all providers of hardware and software.
While this movement has been natural for SaaS providers, there are interesting variations that occur when the same methods are applied to hardware, as well as services tied closely to hardware.
Kevin Christian, presently with Infoblox, is an industry leader with a wide exposure to pricing strategy and operations. He has run the pricing function at several companies who are leaders in their own industry segments. He also founded his own consulting business with the same focus area. Through his long presence in the industry, he has been witness to the transition of the IT infrastructure industry from a licensing regime to a consumption/metered one.
In the following sections we share Kevin’s insight on both his general principles for pricing success as well as a fictionalized case study of NewCo and its transition to consumption pricing.
Organizing the Pricing Program
Per Kevin, at most software companies you can break down a Pricing Program into three big areas, or legs so to say (see Figure 1). The first leg is Strategy which includes programs and policies, financial analysis, market analysis, and competitive analysis. Typically, the changes are planned as short term, say over 6 to 12 months, and long term over the coming 2 to 5 years.
The second area is pricing of specific or individual Transactions. The elements here are what is often referred to as the Deal Desk, or, Commercial Finance. The activities in this area are about approving deep discounts on special, large deals and any non-standard contract terms.
The third area is Tools and Processes. This consists of processes set up for reviewing prices and setting new prices or changing existing prices. Also covered in this group are activities like publishing the price list, setting up skews - it is any system related changes and whatever the quoting process is.
For a model to be successful, it has to have three attributes - Simple, Measurable, and Scalable.
Beyond these essential three, we can list four more attributes to ascertain if the pricing model is effective:
To get an industry wide perspective, Kevin alludes to an imaginary company, that can be called Newco Assets and Operations. This company has offerings in the complete infrastructure space - from networking, IT asset management (ITAM), IT operations and automation (ITOA) to Information Security. This is a fairly typical profile for companies in this space.
Traditional IT Infrastructure and Operations software was deployed on physical systems that were mostly on-premise, where the vendor was far removed from physical access and customers bought licenses to match the specifications of their infrastructure.
In this regime, pricing was anchored by the hardware oriented deployment model, where the customer paid a small price for software running on a smaller machine and a bigger price for the software running on a bigger machine.
However with cloud computing this model is changing and so is NewCo’s Pricing. As shown in Figure 2, the evolution can be described using three dimensions of the problem, each having an impact on packaging:
In NewCo’s case its packaging had grown very complex with the addition of many point products and the company moved to a more simpler packaging structure. In its case it reverted back to a Good, Better, Best style of packaging. (This is something we’ve covered earlier in the book)
It starts with a small (or Good) package which is the minimal viable product that creates value for a customer. The medium (or Better) sized package is designed for the vast majority of customers. Finally, the everything (or Best) package is for the customers with the most well defined needs or the most sophisticated customers who have to have every single feature and function. In this model, the point products and bundles are gone and the company is left with just those three packages.
As part of their own packaging transition, Newco has also arrived at a GBB packaging that covers all of Newco’s offerings consisting of Networks & ITAM, ITOA and Information Security. The customer does not get to pick and choose individual items but gets one complete package.
One problem in this three level packaging is that the Sales people would like to sell an entry or lower level package first and be able to go back later to try and sell the higher one. That is because they like to sell more to existing customers. To grow, the Sales team would take an ITAM customer and sell them ITOA or Information Security. Similarly, if they have an Information Security customer, they sell them ITAM or ITOA, or, if they have an ITOA customer, they sell them ITAM or Information Security. Customers don't buy all three products, initially.
For Newco, how this works out is that they now have their Essentials, Business and Advanced packages. So if you have a customer who bought Essentials, Sales would try to upsell them to Business, and if they had bought Business, Sales would try to upsell them to Advanced. Now, if they have already bought your best version there is no upsell because they are already at the top.
One way to get over this situation is to look at another approach to growth, and that is to look at the pricing metrics or variables in each of these packages. For example, Newco can use the size of the customer’s network in terms of the number of IP addresses. In this case, Newco can capture growth if the customer’s network grows from 10000 IP addresses to, say, 50000, thus scaling five times.
There is a case to be made here for still having a very small number of point products on top of the Good, Better and Best packaging. When a company introduces this type of packaging it works for a year or two. At this point they could start introducing a very small number of very expensive point products. Of course these would have to be very specific high value, expensive to deliver, point products. Because they are expensive to deliver, they would have to be priced high in order to make any money on them. In this way the Good, Better, Best model is still able to accommodate point products.
Taking the example of Newco, they had similar decisions to make across their Networking and ITAM, ITOA, Information Security portfolio. The key decision they had to arrive at was whether and how to have a single usage metric and if they would need different metrics for their three offering groups.
Since the Networking and ITAM offering itself has multiple products lines they could, technically, have different metrics for each of those. These could include the number of DNS Queries or Queries Per Sec (QPS) for DNS services, and for DHCP services it could be the number of Leases or Leases Per Second.
Similarly, the ITAM pricing could be based on the number of assets or the number of IP addresses. The ITOA pricing could be based on the number of devices being managed. The Information Security part could be priced on the number of users or employees.
Now if they used all of those metrics then they could either sell them separately or convert them into 15-16 bundles starting with what would be appropriate for a small attorney’s office and going up to a large Walmart type of facility and then the typical corporate office of a large multi location corporation. For each bundle, they would need to specify a ceiling for the respective usage metric of each of the five product components within that bundle.
So it was apparent that this would become really quite complex and yet it would never fit exactly for any single customer; they are always going to need more of some item and less of some other. And even then they may feel they are overpaying for something.
Even though this arrangement is complex, a number of companies price their offerings like this. One example of this could be Cloudflare. As Figure 3 shows, Cloudflare has a Free plan and then Pro, Business and Enterprise plans. Then if the customer needs more they can take from a set of Add-Ons on top of these plans as depicted in Figure 4.
For Newco it made more sense to have a single metric for everything. Newco needed to not only look at what was theoretically best for them, but also their established sales motion. In Newco’s case they decided to go with a single metric because that was the preference of their Sales and field leadership and that group values simplicity more than anything else. The field leadership wants to do big deals where the customer buys all the products.
In Newco’s case the single metric used became IP addresses for everything, even though IP addresses are not the most natural metric for some of those items. In this case, they would be using a suboptimal metric for some of the items, but the reason they were well advised to do it was to have the simplicity of the same metric for all the items. Using this single metric, they were been able to develop the Essentials, Business and Advanced packages for each of Networking and ITAM, ITOA and Information Security.
At this point Newco has been over the transformation and simplification of packaging that they attempted and the single metric approach to pricing that they liked. Now they have to deal with how they roll this out to customers and their own field teams.
We start with a metric that we wish to use, and in Newco’s case it is the number of IP addresses.
Newco has multiple ways in how they can make the customer pay for what they consume. The first method is that the customer pays for a certain quantity and they cannot go over the ceiling. The second method is that the customer consumes what they need and Newco bills them for the actual arrears based on what they consumed for a period that has been chosen, say per month. That is the method AWS and Azure use. It is also possible to bill the customer in advance for a certain amount but they can go over the ceiling and at the end of the period their ceiling gets resized. This flex method has now become quite popular with the larger cloud vendors and has become a recommended model as well.
Having developed a mature and elegant model Newco still has the challenge of getting it implemented internally. If we go over the process again we have four steps to complete:
In this transformed approach, the Sales teams get compensated for upselling a customer on new functionality, or they get compensated for increasing the quantity that the customer is buying or some similar change. They don't get compensated for anything else because the CSM team has taken on most of the post sales work. In that sense, the CSM team is fully focused on the data being generated out of the selling and consumption.
The key challenge with this approach is that the CSM team cannot be a success unless they have all the data on usage being generated. Having the data ready is critical for the way we want to price now - pick the metric, count it, collect it in a data lake, show it in a dashboard. Unless we can do those four things we cannot get the new pricing model implemented well. This is not a problem that is confined to one company but is prevalent quite widely. The PMM or Pricing leaders often face an uphill climb to influence Product or Engineering in a manner that prioritises this instrumentation for pricing. But it is important to get the new Pricing actually work.