Hosting Is Dead, Part II
This blog is intended to give you detailed market insights on a trend visible since approximately 2017: The large-scale exodus of enterprise customers moving away from hosting providers to hyperscalers.
While the first part laid out the mechanics driving the migration, this blog post focuses on the impact of this tectonic shift on the hosting provider business model, their customers and employees.
Part II: Impacts
So, now we know that hosting providers continue to face pressure coming from the hyperscalers, as defined in the first part of this blog post. So far so good.
The exodus is currently accelerating further. More and more hosting provider customers are terminating their contracts and following their newly identified Cloud First strategies. Maybe prior to the last hosting contract renewal, some were not sure about moving to the cloud – no matter if IaaS or SaaS. This happened typically 3-5 years ago, the industry standard duration for hosting contracts for many years.
As strategies have changed for an increasing number of clients towards Cloud First in the mean time, the exit rate of customers now is exceeding the number of new customers for the hosting providers.
So now the downward spiral kicks in: negative economies of scale.
Negative Economies of Scale
A quick explanation first, so we are on the same page: Economies of scale are a basic economic principle. They are achieved through cost advantages coming from an increasing amount of output volume. Consider fixed costs, e. g. machine depreciation or salaries, being spread over an increased output volume. The result is a lower cost per unit.
If we consider the opposite effect, fixed costs spreading over less output volume, prices will increase. This is unwanted behavior, sure. And this is the case for the hosting business model, which we will analyze now more closely.
Pressure on the Business Model
As outlined above, more hosting provider customers are leaving the hosting market than new customers enter. For the hosting providers there are two effects creating pressure on their existing business model: increasing costs coming from negative economies of scale and fierce competition tackling the prices and thus, margins.
With the decreasing number of customers, negative economies of scale apply:
The systems needed to provide storage, backup, network etc. must be scaled down to avoid additional costs from unused capacity. Smaller systems need to be integrated into the datacenters which create additional effort and lower economies of scale by design. Costs per unit increase further. A vicious circle.
The consequence is an increasing cost per unit, which further accelerates with every customer lost to the hyperscalers.
The increase of unit prices does not come at an ideal moment. With every hosting customer leaving the hosting space, this exact market continues to shrink. And this further decreases the chances for the remaining hosting providers to win new customers or retain 0existing ones. In the end, the hosting market is a very mature market, that has developed over the last 30 years. Mature and saturated markets tend to be dominated by price. And for the hosting providers this means fierce competition over the remaining hosting provider customer potential.
On the one hand, hosting providers need to increase prices to keep their margin at a safe level. This is based on the negative economies of scale, so increasing costs. On the other hand, the market situation getting more and more competitive forces the hosting providers to reduce prices to win customers adding up to the deserted datacenter racks. For the hosting CFO ‘s this is the definition of nightmare.
War for (Cloud) Talents
Since many years, employers are fighting for the best talents. This is based on increasing demand for highly skilled knowledge workers. On the supply side though, several factors reduce the amount of available top talents, like demographics, urbanization and so forth. McKinsey created the term “War for Talents” in the study “Attracting and retaining the right talent” published in 1997. This denomination is now used globally.
IT Labor Market
For the talents in the IT space, the labor market mismatch is even more drastic. Several countries list hundreds of thousands of open positions that can not be filled. For the German IT market for example, the average hiring period in 2021 was seven months.
Where does this come from? Digitization projects and hyperscaler cloud growth are multi-year megatrends draining the available talent pools. Attractive brands in the cloud, like Google, Microsoft, Apple etc., are typically first on potential employees’ lists and they empty the market for hosting providers and on-premise customers alike.
Top Talent Leaves First
Unfortunately, the pressure on the business model is transparent to the hosting providers’ employees as well. They notice their clients exiting the datacenters. And they ask themselves what their future will look like. Maybe they should start to listen to the headhunters knocking at their doors continuously and talking about cloud.
So, what now happens is that the most experienced operations specialists leave the hosting provider first. This leaves the hosting operations without their last level, who typically fix the most pressing or most difficult problems. On top of that, new talent is slow to hire and reluctant to join a hosting provider, if they can become a „Googler“ instead. This leaves the remaining operations teams with increasing workload – and stress.
To round this up, there is a psychological effect. Less experienced operations specialists start to ask themselves, if they should leave the sinking ship as well.
Another vicious circle.
Impact on Hosting Customers
The commercial and resource changes described above impact both existing and new customers.
New potential customers might notice pricing outliers during tenders. This is due to the negative economies of scale outlined above. They lead to obvious price increases to save margin levels. Or they lead to price reductions in competitive situations.
Apart from this, new customers might be told, that some reference customers are not available anymore – probably because the latter terminated their contracts.
For existing customers on the other hand, the picture is more elaborate by design. Let us start with the commercial changes. The hosting providers have to protect their margin against negative economies of scale.
This can be seen first with consumption-based items likes storage units. Higher prices will create backlash, as hosting providers are not used to price increases. Additional demand, so-called change requests, tend to be more subtle. Higher unit prices for storage and compute are mixed with managed services, which makes it not very transparent and hard to compare.
A totally different story is the preparation of a contract renewal. Given the nature of a renewal, any price change is very transparent: You re-purchase the same at a – presumably – lower price.
Loyal customers will probably feel betrayed and react all guns blazing.
Data Center Consolidations
Apart from the direct costs addressed in the price increases section above, a typical area for cost cutting on the indirect side is the consolidation of data centers. Instead of having three datacenters at partial capacity e.g., they concentrate their regional customers in one datacenter to improve economies of scale. Already today, many hosting providers force customers to move to “more modern” datacenter locations.
Unfortunately, this gives customers the chance to apply extraordinary termination rights or to rethink their deployment strategy in general. The hosting providers are fully aware of the risks but need to act quickly.
Changes in service quality apply to the operations team customers interact with on a daily basis. Very senior experts will be replaced by new colleagues. They are the first to leave, as outlined above in the Top Talent Leaves First section. And the replacements will probably not be as experienced as the ones known to the customer over a longer period of time. If the working relation was personal, the customers will probably be informed about the new employer the experts have moved to. Customers won’t be happy, but they do understand change is needed every now and then.
What they will not understand is why requests start to take longer and longer over time. There are two reasons for that: First, less experienced staff will need more time to solve the problems at hand. Second, the resource gap grows bigger and bigger. New talent is difficult to hire if you are a hosting provider. This takes time and is costly as well. Please refer to the IT Labor Market analysis above.
Due to staff rotation and resource scarcity the error rate will certainly increase. This is a very sensitive area for the hosting customers, as business critical enterprise applications are at stake. Should the reason for an outage be identified as human error and the weak spot as one of the new team members, hell breaks loose for the hosting provider.
Already alerted by price increases, staff leaving the company and extended ticket resolution times, customers will start to re-evaluate their deployment strategy. Or prepone decisions with regards to the cloud. So, if they did not decide to onboard the Cloud First train, it is now that this happens.
Summary and Outlook
Over the last 10 years, the perception of cloud from the IT executives’ point of view has changed. As have market conditions. This has led to infrastructure workloads being transferred from hosting providers to the hyperscalers.
Unfortunately for hosting, this has consequences. Negative economies of scale increase unit prices for the hosting providers, who find themselves already in a highly competitive market. The catch-22 situation is to raise costs and lose customers more quickly or to reduce costs and burn capital.
Customers are leaving anyway and this is noticed by the hosting operations teams as well. Top talent will move to the hyperscalers first, being followed by more and more former colleagues. With the War for Talents raging fiercely in IT, replacing the talent needed for continuing customer operations is no easy task for hosting providers. And now, customers start seeing changes in service quality as well, driving more customers away from hosting.
“(…) the species that survives is the one that is able best to adapt and adjust to the changing environment in which it finds itself.”
— “Origin of Species”, Charles Darwin (1859)
After 30 years, which is a very long time in information technology, the hyperscaler comet has started to wipe out the dinosaurs. The predominant species are now IaaS and SaaS. Some of the dinosaurs will find a way to adapt, e. g. become system integrators synergizing with the hyperscalers.
And some will continue to exist in their specific niche.
In a nutshell: The era of hosting is over.
More in-depth information on hosting and cloud deployments can be found here:
Follow me, so you do not miss any of my future blog posts!