Expand your Cloud Center of Excellence to the whole organization and become a cloud-first business

Published in

Google Cloud - Community

7 min readJun 28, 2019

tl;dr There are plenty of resources explaining what a Cloud Center of Excellence is and how to set it up (see further reading). I will share some guidance on what comes next. Which is, expanding what was learned and built in your CCoE, to the whole organization. Without becoming a bottleneck.

Introduction

Given today’s challenges in IT and the vast amount of consultants, vendors and internal advocates, it is highly likely you already have a Cloud Center of Excellence (CCoE). If not, maybe you have one, or more teams who do the same thing but sail under a different flag like Cloud Competence Center, Cloud Tribe, DevOps Team, Cloud Magicians or just “the folks who do this new thing called cloud”. Although I’d like to challenge, if cloud should be considered being new.

Your CCoE is a mix of engineers, operations / SRE experts, project managers or Scrum Masters (which is not the same thing, of course). Leadership, especially a thought one is also part of it. Not to forget marketing and communications as well as change management. All of them form a great team, having embraced the concept of cloud.

You have achieved first results, ready for production. Or even run a live e-commerce platform, rolled out a machine learning solution to identify fraud, or enabled your customer contact center with sentiment analysis.

And — I’m pretty sure — you identified growth issues, bottlenecks and difficulties in rolling out, what was learned in the CCoE, to your wider organization.

I will help to identify pitfalls a priori and share ideas on how to avoid them. Moreover, I will propose, how to scale your CCoE’s competence and expand from a cloud island, to every shore in your organization. And become truly cloud-first.

Common restrictions a CCoE faces and how to avoid them

CCoE is slowed down by governance, security or budgeting

This is a tough one, with books to fill. In a nutshell, provide senior executive sponsorship to the CCoE, helping to obtain required process exceptions (e.g. financial planning, forecasting, hiring). At the same time, foster a cultural change towards more risk taking and measure results in impact instead of quantity. For security, establish a clear set of rules on what can go to the Cloud, what can’t and which are the security requirements. Make sure those rules are based on actual requirements (e.g. regulations, industry best practices), rather than on gut feelings or lack of understanding Cloud.

CCoE is perceived as just another supplier management layer

While the CCoE knows Cloud suppliers’ products, processes and support procedures by definition, make sure not to off-load highly specialized tasks like contract & supplier management, troubleshooting or support to this team. Instead quickly scale those capabilities into departments, already responsible for for such tasks in “legacy” environments. In addition, enable your CCoE to uphold and enforce defined Cloud standards.

CCoE struggles with implementing multi-cloud everywhere

I consider multi-cloud as a given in large enterprises. Choice is a good thing. Still every supplier you add to the portfolio will increase complexity. With regards to hyperscalers (Alibaba, AWS, Google and Microsoft), try to start with one. Scale only after you understand the concept and mechanisms of Cloud well and have a business reason (e.g. unique feature, performance, availability) or regulatory requirements. Don’t try to make each of your services available on two or more hyperscalers (except services like Kubernetes or Anthos). The efforts of doing so and incompatibility of features would slow you down and bind unnecessary development, testing and integration cycles. Not to speak of trying to back-charge Cloud cost to individual services.

CCoE tries to avoid vendor lock-in

Some services (e.g. Kubernetes, open source-based databases) are very portable. Yet others are more difficult. I see no difference to previous technology or software purchasing decisions. Make sure to make a conscious, fact based decision when deciding for a potential lock-in. Don’t over-engineer, or try to build abstraction layers beyond what is available as open source (e.g. Terraform). Such efforts rarely pay off, if not backed by a large community.

CCoE focuses on individual building blocks (e.g. architecture, CI/CD) instead of end-2-end Cloud

Avoid by staffing diverse backgrounds (developers, architects, operations, business) and define 360 degree OKRs with transparent ways of measuring. Make sure an end-2-end perspective is baked into the CCoEs OKRs, which can then be broken down to individual responsibilities. It is vital for staff to identify themselves with their individual, but also overall OKRs.

CCoE is not visible in the organization

Create awareness by giving the CCoE seats in standardization and decision bodies like the architecture board. Also encourage them to speak in internal events related to IT and business transformation. In addition use internal blogs and newsletters, or whatever your organisation is using there. Ensure that the CCoE is creating products with business impact. Visibility will come from there.

CCoE does not attract the envisioned talent

Internal: Have a clear career path for people joining, but also leaving the CCoE. Allow for moves sideways and encourage seasoned leaders managing people to make a “downward” step, by temporarily leaving their people management responsibilities, while providing their know-how and full attention to the CCoE.

External: Speak on external events and publish content your CCoE developed publicly. Provide perks and adequate salary required to attract required talent.

Both: Rigorously check your recruiting process and tailor to your CCoE’s needs. Adjust depending on the maturity of your CCoE and hire only new staff that is exceptionally good in one or more specific areas which you need.

Ways to expand a Cloud Center of Excellence (CCoE) to your wider organization

Once your CCoE is successfully established and delivering first results, you want to scale those learnings and assets to your whole organization. Here are some ideas on how to do that.

Tell a business story

Cloud doesn’t exist for the sake of technology. It is a vital tool for the digital transformation of your business. This should be reflected in the strategy and the stories you tell. Always have a reference to the business impact when talking Cloud. Users need to see the benefits for their work and business problems. Everything else will follow.

Encourage and allow for independence

Most of the teams adopting cloud and building on top of the ideas and assets of your CCoE will come back to the CCoE. They will ask for help, guidance, reviews and approvals. Avoid the latter at all cost. Design the decision process for scale, not to become a bottleneck.

Also provide guiding principles to teams, so they don’t need help from the CCoE in standard situations. Offer technical and non-technical training capacity, as well as budget suitable for the amount of staff requiring training. Include different types of training like self-study, mentoring, classroom, labs or Hackathons. Create learning paths, which will enable people and give them the required confidence for independence.

Trust in the foundations you built. Your organization and staff will be able to sort it out. Encourage them to do so.

Evangelize

Encourage CCoE staff to be on stage (internal and external) to grow awareness and build networks across the organization, make others wanting to try new technologies and solutions and also attract new talent.

Make sure to work on publications to spread the word. Leaders and sponsors in your organization need to support those efforts and participate whenever they can to show buy-in and provide confidence.

Disseminate

Re-assign CCoE veterans to other units in your company, either as leaders, experts or individual contributors. Make sure this move is aligned with expectations of a particular person and the career path communicated to your CCoE staff. Such individual tasks could be:

Asking a Cloud Analytics expert to transfer to your finance department to help building more transparent and accurate reporting.
Having somebody with deep Machine Learning (ML) expertise support your manufacturing line on predictive maintenance and teach them what kind of data is required, to produce good models and predictions.

Use those veterans as advocates, evangelists but also as a feedback loop. They suddenly experience Cloud from a business or IT perspective instead of a CCoE one and will definitely come up with first-hand ideas on ways to improve. In addition allow for temporary rotations of other staff into the CCoE to learn and disseminate once they rotate back.

Grow Cloud footprint

Based on your digital transformation strategy, define a roadmap of successive, additional cloud workloads and tie organization-wide OKRs to them, to get things moving. Break down to individual OKRs to ensure individual ownership. Don’t start with huge complicated all-in programme efforts requiring months of planning, vendor selection and huge teams to staff. Pick easy wins, learn and grow from there. Cloud adoption beyond single business problems and PoCs will lead to bigger impact, freeing up capacity through automation and result in more acceptance. In essence, growing your Cloud footprint will be one of the biggest drivers for CCoE know-how transfer.

I am sure there is even more which worked out for you. Or bits and pieces which didn’t go well. Please let me know!

Next steps

Pick two areas for improvement and define an approach on how to tackle them. Get management buy-in and sponsorship. Then iterate. Definition and decision should be in the days than in the weeks.
Agree with your CCoE staff (e.g. in the next team meeting) on three public events they are going to speak at. Make sure the speakers are / get trained well and rehearse internally.
Revise your CCoE recruiting approach and sharpen the career paths.