Let’s be real: talking about tech migration isn’t exactly the most thrilling topic. But you know what is exciting? Saving millions, simplifying operations and future-proofing your business.
In this episode, we cut through the noise and break down why media and entertainment companies (and others) can’t afford to sit still on their legacy systems. From hidden costs and security nightmares to missed innovation opportunities, we expose the risks of delaying modernisation.
Inside the episode
- Big bang vs. phased migration: what works (and what will destroy your budget)
- AI in cloud migration: hype or actual game-changer?
- Why simplification sounds boring but is your biggest competitive advantage
- The brutal truth: not every system deserves to survive, so be ruthless
If you think you can just “lift and shift” your legacy tech and call it a day, think again. We’re here to challenge conventional thinking and give you a no-nonsense roadmap to smarter migration.
Simon McGrath and Alex Borland are joined by Chris Birkinshaw, Technology Principal, and Rod Fairweather, who had joined Merapar as a broadcast architect that same week. Alex opens with the themes coming up across the industry: financial pressure, with advertising and subscription revenues down and global OTT players moving in, and a shared need to modernise and simplify. Chris describes clients that are mature in the cloud but still run large on-premises systems they have avoided touching.
The conversation then turns to how to migrate: why replicating the old system is the costly route, why a feature slice delivers value sooner than a component-by-component migration, and why big bang should be the last resort. Rod brings the operations view, Chris covers DevOps and testing, and the panel asks where AI helps. It closes with Chris’s advice to a broadcaster’s CTO met in a lift.
Highlights
- 07:56 “The biggest pitfall that I see is that people don’t engage with product.” Chris Birkinshaw
- 13:53 Chris Birkinshaw on slicing by feature, not by component: in a sports data architecture, migrate Premier League football first and the most important sport is protected.
- 18:06 “Our housekeeping is very, very poor as an industry.” Rod Fairweather
- 24:36 “AI is not magic. You can’t just go into it and say, hey, build me a new system to replace this legacy system.”
- 28:30 “You need to be brutal. You have to turn some stuff off and not replace it.” Chris Birkinshaw
Key questions answered
Why are broadcasters tackling their legacy systems now?
Economics, and a ticking clock. Alex Borland points to falling advertising and subscription revenues and the threat from global OTT players. Chris Birkinshaw sees outstanding security exceptions, systems on unsupported hardware, operating systems or databases, and vendors end-of-lifing products. Add in-house applications built 15 or 20 years ago and poorly documented, and the people who know them retiring. The episode sums it up as cost, risk and opportunity.
Should a migration replicate what the old system does?
Usually not. Chris warns against a technology-led swap of one system for another, which means first working out what the old one does, then reproducing assumptions shaped by its constraints. Instead, go back to business stakeholders and users, build a prioritised list, and accept that some features will be dropped: there is probably no business case for all of them. Any technology change also needs organisation and process change around it.
Big bang or phased?
Phased, in thin end-to-end slices. A component-based migration delivers nothing until the end, and a programme that disappears for a year means building every new feature into two systems. The advice is to get the happy path working before the edge cases, treat big bang as a last resort, and ideally ship something the old platform could not do, so people outside the tech team see the value.
Where does AI help in a migration?
In the supporting work. It can help analyse existing data sets and documents and find the gaps, and it helps developers write code, which still needs human review and has to stay maintainable. It is particularly useful for adding tests retrospectively: integration tests around a payment API, for instance, so the interface contract holds through the migration. Chris’s wider point on testing is to shift it left, not leave it all to the end.
What should a broadcaster’s CTO do first?
Chris’s answer: be brutal, turn some things off without replacing them, avoid analysis paralysis, and start on something with business value, expanding as you go. Rod adds that speed matters once a company has decided to act, and that an outside team which has already seen different solutions, from large workflow engines to Step Functions in AWS, helps get things done in time.
Speakers
- Chris Birkinshaw, Technology Principal, Merapar
- Rod Fairweather, broadcast architect, Merapar
- Simon McGrath, Merapar (host)
- Alex Borland, VP Strategy & Business Development, Merapar (host)
