At the PoweredByCloud conference earlier this week in London, around 170 CEOs, CTOs, analysts, lawyers, universities and users from the cloud computing community met for three days discussing developments in cloud: its adoption, barriers, business models, success stories and legal issues.
Interestingly a lot of the talk actually centred on getting out of the cloud. It appears that among those users and potential users that have done their homework, the business case for cloud is satisfactory and the use cases so far disseminated encouraging. However along with the security, privacy, reliability, compliance and similar concerns that we have been talking about since the start of the field, one of the major barriers to entry for many users is that once in a given cloud, is it possible to get out again?
The scope of this question ranges from the concerns regarding provider lock-in to emergency withdrawal in the case of a breakdown of relations with the provider. The view from many of the smaller players and the “cloudarati”, as someone has coined them, is that cloud is a commodity and as such it should completely painless to move data and applications between competing clouds.
However, these are also barriers we see in a number of other sectors without such great concerns, and vendor lock in is rarely (if ever?) absolute. Really we should be talking of switching costs, a term that emphasises that it we are dealing with only half of the equation, because countering the cost is the benefit. When the benefit outweighs the cost then we are prepared to make the change. It only becomes effectively absolute when that cost is insurmountable, irrespective of the potential benefits. But switching costs are something we all accept all the time. When you change your telephone provider you may be charged a contract termination charge by your old provider, or even a connection cost by your new provider.
Let’s take another example. When it comes to the banks and mortgage lenders it is usual to find a myriad of different terms and conditions relating not just to the interest rate but also the costs of setting up the account, the cost of paying it off early, different if it is paid in full or in part, and so on. Do these deter people? Clearly they do not, despite much of them being overtly groundless charges. One has to look at the entire package. It is actually similar to cloud provision in several ways: it is a potentially long term decision, you must trust the provider (at least somewhat), banking is a commodity and cloud is heading for one, and you have relatively little loyalty to one provider or another. And yet people are happy to accept mortgages with explicit switching costs as part of the contract, when we know there is no basis for the charge, so why not for cloud?
True, when it comes to mortgages the numbers are fairly easy to compute. You know the switching cost (say 1%) of closing the mortgage to go to competitor. You know the competitor rates and you can calculate how long it takes to amortize the switching cost. You know the time remaining on the mortgage. You make the decision based on the magnitude of the eventual saving. In cloud the situation is more blurry: the time it takes to set up the new infrastructure is only an estimate as are the exact benefits from the new provider. Nonetheless, any decision to change provider is based on both sides of the same equation. Cost and benefit.
Essentially, then, the attitude of the cloud providers to deliberately building switching costs into their business model should be based on business principles. If you offer long term savings you may be able to impose higher switching costs. As you are trading a commodity, you must look at the overall benefit and cost you provide. Indeed having switching costs may well afford a buffer for fending off rivals on a temporary basis. This may actually be beneficial to the long term profitability of the market, likely to see profit margins eroded in a price war. By the same token, other providers, particularly the smaller, newer or otherwise more expensive players may need to eliminate switching costs entirely to attract clients in order to get a foothold.
However, the situation is perhaps not straight cut. As well as the economics of the situation, there is also the inertia of the situation. In banking, the first mortgage lender you approach is your existing bank. You don’t swap provider over a trivial amount. Cloud will be the same. Switching costs also include the investment in time and the consequent disruption the change causes.
Currently the market is young and even Amazon, the assumed dominant player does not have it sewn up, not by a long stretch. Most providers have some switching costs, different APIs and configurations. Providers are still competing as much on differentials as on price. However, as the market evolves towards full commoditization, the companies will become less differentiated on offering and on price. Now is the time for those with the stronger position to start a land grab. If they can keep those clients during the journey to commoditization in the next couple of years, even low switching costs may become enough to mark the difference between those still trading in 2020 and those that are not.
So in conclusion, providers must know their switching costs. Their customers and potential customers certainly will. They are not an absolute barrier to entry, merely a hurdle, a fee. Providers should attempt to minimise them for new customers and deliberately assign their magnitude for exiting customers, high or low, as part of a larger strategy based not just on today’s market, but also tomorrow’s.