Agile end-to-end

Agile end-to-end

This is the sixth piece in a series of blogs by Phuah Aik Chong, Astro Chief Technology Officer

In some of my past articles, I have covered the transformation journey we have embarked on, focused quite a bit on Lean Innovation techniques, the adoption of Cloud, active talent management and becoming a data-driven enterprise – all contributing to our “digital reinvention”.

This article looks at why and how we are adopting Agile to transform not only the way we develop products and services, but how we constantly stay aligned with our customers’ desires and also how we remove internal organisational barriers and apply lessons learnt along the way.

Over the past few months, as part of our journey in adopting Agile, we realized that Agile is not just about software development but about being Agile end-to-end. To many people, Agile is simply an iterative software development approach which delivers value earlier than Waterfall, with lower risk, better quality and helps keep development teams and the business aligned, because they work together in the same teams doing things in small iterations. While this alone is a major improvement over traditional Waterfall approaches, there is more... Agile is a mindset change that should influence our overall organisation culture. Agile is not going to be successful if it only takes place in the software engineering team.

Why is Agile important to Astro?
The media industry is probably one of the most disrupted industries in this digital revolution. Product lifecycles are getting shorter, customers are overwhelmed by choices and new players can quickly disrupt established leaders. To survive in this digital era, Astro has to accelerate and stay future proofed by delighting our customers through continuous innovation.

The first reason for adopting Agile is that it improves our focus on Customer Value. By constantly prioritising product or service features based on the value perceived by our customers, we ensure that our products and services stay aligned with their expectations. We measure the value by observing customer behaviours when they are exposed to new features, not using our opinion or even their feedback. If experiments show that the perceived value is low, then the decision to reprioritise features or even cancel them becomes natural. This is why experiments introduced by Lean techniques fit well with Agile. For example, we have seen an increase in video consumption from continuously experimenting with how Video On Demand content is being surfaced and recommended to our customers. This is both on our connected set top box and the Astro Go video streaming product, and we have also been experimenting with different methods for users to sign up to consume our digital products in order to ensure we provide the most frictionless registration journey.

Secondly, as Agile encourages frequent releases of smaller (therefore less complex) batches of work, as opposed to large monolithic releases, the level of risk of each release has reduced and the amount of rework has come down significantly. Moreover, since each Agile iterations include design, build and test activities, defects are identified earlier and the overall quality is improved. For products that have adopted agile development, we are now able to demonstrate them to stakeholders after every 2-week sprint cycle, then have it quickly tested & fixed (if there are any bugs) instead of waiting for multiple months before seeing something at the end of a long process. The DevOps techniques such as continuous integration/continuous deployment and automated testing further enhance the velocity of the engineering team.

Once teams are focusing their efforts on proven customer value and focusing on smaller batches of work, changing course when necessary becomes much easier and this is why Agile enables us to constantly and quickly re-adapt ourselves to changing conditions.

Why is Agile a challenge to Astro?
Moving to Agile has not been that straight forward for us; the fact is, we have been in this business for over 20 years and there are various established processes, legacy technology stacks and skillset constraints that we have had to overcome in order to shift gear.

Changes of this magnitude take time and they are not easy. They require a genuine intent and continuous focus from the senior leadership team. We are fortunate in Astro that we have had the full support of our senior management to drive this initiative and help us to ensure we could transform Astro into an agile, nimble and innovative company.

We now need to shift our business and project management mindset from being timeline focused to customer value focused. Instead of simply holding delivery dates as the main measure of success for a product launch, we need to move towards making sure we deliver the right experience and value to our customers before releasing the product or services to the market. At the end of the day, it is about the customer, not us!

There is often a misconception that Agile will allow us to deliver the same scope of project faster than waterfall. Agile doesn’t produce miracles though, it’s just a smarter way to prioritise and organise work. Therefore, if a large monolithic team was needed to develop a given product, introducing Agile won’t magically enable the same product to be developed by a small Agile team as several smaller teams will be needed instead. Their cooperation and overall alignment with the Astro strategy is key to our success. To achieve this, teams should be “loosely coupled” which means the way they interact should be based on documented interfaces or engagement models and the work should be distributed among teams in a way that will not create bottlenecks if one team is delayed for some reason. Each team should be skilled and empowered to cover the end to end activities required to create customer outcomes without depending on the rest of the organisation.

Being pragmatic (and patient) during periods of change is critical. It’s not about imposing Agile to every project, all the time. For some projects, traditional Waterfall approaches are still valid, especially when scope, requirements and technologies are well understood and stable. Even for projects that have been identified to be an Agile candidate, it may still take some time to shift gear.

So how are we actually doing?
To adopt these principles, we are reorganising ourselves to focus our energy on a few important initiatives. This will drive better customer experience than spreading our resources too thinly over multiple products and services, which ultimately would not benefit our customers. The learnings from these pilot projects can then be applied across other products/services.

We are dedicating self-organising and cross-functional teams to these initiatives that we call “Bubbles”. Each Bubble will have all the resources it needs to deliver a product, including both business stakeholders and a complete dedicated engineering team (known as a Service Team) from our Technology Centre of Excellence (COE). The Service team will have all necessary skillsets required to be able to perform an end-to-end engineering function from product management, project management, UI/UX design, architecting, coding, QC, test automation and 24/7 support etc. By doing so, we will empower our best talents to delight our customers and minimise dependencies on a bureaucratic, top-down approach. These teams will be autonomous and accountable for delivering the right customer outcomes.

We have been running Agile projects for some time now and we continuously assess the results to further streamline subsequent initiatives:

  • We found that we needed better alignment between the KPIs of business and development teams to ensure that all team members are focusing on the same priorities
  • We needed to improve the “Agile proficiency” of our development teams by standardising on roles, toolsets and the language they use
  • Another finding from these first few Agile projects was the need to “connect” the tactical team-level priorities to the overall Astro strategy and to put in place a governance structure to guide the autonomous team to stay on track

These are some of the key challenges we are solving to ensure that the “Bubbles” continue to progress well and achieve some of our strategic big bets.

Encouraging the Agile practices
We want to give every Astrolite the opportunity to be part of this endeavour. As mentioned in one of my previous articles, our current teams are enrolled in several training programs around Innovation (Certified Innovator Programme) and new Cloud related technologies (Certified Technologist Programme).

We have also initiated a similar program for Agile through the “Agile roadshow” which gives an initial level of awareness to a large number of Astrolites. We have helped 25 of our employees to achieve their Scrum Master certification. Additionally, in partnership with AWS Professional Services, we ran a 2-day “Business Agile” training for our Business and Product Owners, during which we introduced useful techniques such as Amazon’s Working Backwards, Agile estimation and Lean Product Management.

These activities raised the level of understanding and interest around Agile in our teams and we now plan to accelerate the adoption of Agile across Astro through a bigger organisation-wide awareness programme together with our human resource team. This applies to teams at all levels, from the project level to senior leadership. We have established an Agile community to encourage the exchange of best practices and mutual support across teams. We will support our teams by providing them with coaching and training when needed and we will develop our own (lightweight) Agile standards to ensure consistency of practice throughout our organisation.

More importantly, we will use Agile daily for actual product and service development. Only by practicing and making mistakes we will learn to get better together. Given the progress we are seeing, I am confident that Agile practices will very soon become the norm in Astro.

Click here to read previous blog entries.