In recent years, cloud services have matured to the point where they are now considered part of the sourcing landscape. But apart from the small start-ups that have no legacy IT, no enterprise customer has the luxury of being able to get all of their needs from a single cloud provider – and nor should they as this concentrates risk in one party.
Similarly, cloud providers can’t deliver everything required. The market is not yet capable of addressing all requirements or allaying all the concerns that typical organisations might have. Limitations include the ability to provide fully regulated solutions in all geographies, lock-in to provider-specific technology or commoditised one-size-fits-all service levels.
We see the organisations we work with looking to progressively adopt cloud services and incorporate cloud thinking and tools into their own IT teams – with the aim of implementing a customised hybrid cloud approach that is at the heart of all enterprise solutions for the next five years. To achieve this, engaging a complex mix of parties is essential. For instance, a traditional IT outsourcing service integrator might lead with their own ‘white labelled’ cloud services, delivering operating systems to host legacy applications integrated with functionality from the specialist cloud software as a service (SaaS) providers.
The starting point is to recognise that there is a tension between what the lead systems integrator would ideally like to do – provide a solution that guarantees them a stable revenue stream – and what you want to do – deploy flexibly and scale up to meet demand.
Drawing on our experience of negotiating with service providers to implement and run hybrid cloud solutions for our clients, we think it is best to structure the deal to emphasise flexibility without lock-in while acknowledging that the provider will want to guarantee a level of commitment. Below we outline what this means in practice.
Identify services that are modular and aligned to the business function
So rather than considering services as a set of technology stacks (eg servers, storage and network), consider it as services to support the HR function that comprises of infrastructure components, platform components and application components. If services are then priced according to these components, it becomes easier to compare internal solutions with public cloud SaaS alternatives as well. It can also help make the case for migrating virtual platforms onto public cloud infrastructure within the lifetime of current contracts.
Independently source some of the strategic service components
Where possible you should avoid using an outsourced service provider’s proprietary solution to prevent ‘lock in’. This could include the management of the configuration management database or the service desk ticketing tool. For example, Astra Zeneca moved from using several service desk solutions all run by different teams to a common ServiceNow solution implemented by Astra Zeneca. This provides a single version of the truth which prevents incidents getting lost in the handover between different teams.
Define and price the service catalogue
It’s also important to identify the service catalogue requests from day one and their anticipated volumes. There should also be the assumption that development and operation is a continuous process and that automated tooling is used by the service provider (as it would be with a pure cloud provider) to manage the service request process. Then if the tooling isn’t there on day one the service provider will be incentivised to implement it.
Reward measurable, continuous improvement
Contract lengths are getting shorter and this is not just in the world of cloud. But the extensive use of cloud solutions means service providers have less opportunity to manage their margins on short contracts. Service providers are well aware of this, so consider what conditions should be met to allow for a guaranteed extension of a year to the base term. The incentive to achieve defined service improvements such as a specific percentage increase in infrastructure utilisation, or a reduction in the total number of high severity incidents to automatically guarantee a contract extension can be more powerful than relying on rebates to discourage poor performance. These options don’t exist with the commodity service providers, but the mixed solution of traditional systems integrator plus cloud presents this opportunity.
Be realistic about the minimum volumes of work
Flexibility is undoubtedly important but consider what you are unlikely to deploy to the public cloud over the term of the contract, and how much development activity you are really going to need support for. This is what will drive the minimum volumes a service provider will want to know is theirs and what will keep them interested in working with you. You may need to model your workload as a mix of consistent baseline and spikey demand, for example at month- or year-end, making sure you only pay for flexibility for those workloads actually needing it. This is where the mix of private and public clouds in a hybrid solution can make or break the economics of the overall service being delivered to the business.
All organisations are unique, and no two journeys to a hybrid cloud solution will be identical – but they will all face similar challenges. It is only by balancing commercial and technical considerations, while understanding how risks evolve when services are no longer delivered from a traditionally delivered set of infrastructure resources, that the full benefits of implementing a hybrid cloud can be realised.