A local government GDS might work - but only on top of standard technology

Broken delivery model

First, central government has deep pockets and can continually hire contractors to support everything it builds. If it can’t find contractors, it can outsource to a third party. This couldn’t be achieved among 383 locally focused organisations - how would they all agree? - without radical reform to a fundamentally broken delivery model. Think of an imaginary Tesco plc that indulged each of its 4,169 stores with its own separate back office - the IT department would end up maintaining 4,169 IT stacks, albeit with a few common components.

There is undoubtedly a need for a local government GDS (LG-GDS) but it has, surely, to be one that operates on top of a standardised technology foundation, while also addressing the unique challenges of local government, including resource constraints and the necessity for interoperable, sustainable solutions.

From the outset however, we should be clear-eyed that any LG-GDS would need to prioritise hundreds, if not thousands, of local needs. This would be a huge undertaking in itself.

Second, do we really have the skills for councils to integrate all the open software apps that might be built by LG-GDS locally, over and again into all those silos? Think of all those imaginary Tesco stores.

Open source software has and always will be the key to innovation - however, what happens when all those talented engineers, designers, testers and product folks eventually prioritise their careers and families and get better-paid jobs at the likes of AWS, Google or Microsoft?

Third, there’s the huge issue of tech debt. Although open software is free, the ongoing agency fees, upgrade and maintenance headaches are not - look at the pace with which PHP alone is being upgraded.

Everyday challenges

Then there are the everyday challenges - Cyber Essentials accreditation failures, and digital teams struggling to upgrade Drupal 8 to Drupal 9, for example. On top of this, leaders’ appetites for generative AI and sophisticated automations and integrations add further builds, fixes, releases, documentation, maintenance and testing on top of all the Drupal and associated tech stack upgrades. If a few people leave, everything instantly becomes brand-new legacy and a massive cyber security challenge.

This stuff has happened before. No one will remember DotP, a hopeless and wasteful attempt at building a centralised government content management system (CMS). Then there was APLAWS+, a “local government CMS made by local government, for local government”. (As an aside, I don’t like this “us-and-them” phrase, which feels sectorally isolated and adversarial in an era of global technology and standards, not to mention its whiff of complacency).

Although the idea of an LG-GDS is great, it should be underpinned with a fundamental reset in the delivery of local services using a set of common standards

After gaining 40 or so large local councils, APLAWS+ was expensively abandoned because of mushrooming maintenance costs. If you Google APLAWS+ now, you’re as likely to find cat food. The LocalGov Drupal initiative, as a similar venture, will never solve the problem of needing developers to support it, leading to a dependency on digital services agencies - which, lacking their own intellectual property, are in the game to sell developers by the hour.

Couldn’t all that effort be spent elsewhere, solving harder problems, like social housing, adult social care or special education needs? These services are in dire need of platform re-thinking and proper service design.

Massive opportunity

With the massive opportunity that AI offers, surely there are more impactful things to focus attention on? Instead of building the web publishing world again and again, LG-GDS could, for example, curate and support a suite of AI tools with well thought-out principles, data security guidelines and ethical use policies. A centralised cyber testing suite and/or standards could be created - specific for a well-defined local government website tech stack.

So what’s to be done? Buy common software?

With no strategy (looking at you, Department for Levelling Up, Housing and Communities…) our 383 local authorities would surely risk too much by placing all their eggs in a few baskets. However, if we created common standards for interoperability, we would enable a suite of tools that can be interchanged.

The UK once had a local e-government standards body, called LeGSB, which governed a public sector taxonomy - called LGCL and then LGNL and Integrated Public Sector Vocabulary (IPSV) - and also a framework for interoperability, the e-Government Interoperability Framework (eGIF). We now have all that morphed into schemas and standards managed by the Local Government Association.

Those early standards were adopted by software companies, such as Verint, Jadu, GovMetric and others, which have grown into market leaders in local government. Procurement by councils stipulated that standards like IPSV and eGIF must be out of the box and so common publishing platforms emerged. Those same platforms now power most of local government. Publishing, forms, case management and customer experience tools are all commoditised into standards-based platforms and are, to all intents and purposes - done and available as standardised commodity products easily procured on G-Cloud.

Fundamental reset

Although the idea of an LG-GDS is great, for these reasons it should be underpinned with a fundamental reset in the delivery of local services using a set of common standards – which means digital delivery becomes a configuration and not a product development process. We already have products.

A combination of shared digital infrastructure and local innovation can lead to services that are more responsive, efficient, and aligned with users’ needs. An LG-GDS for shared digital services could include mechanisms for managing tech debt effectively as well as promoting and enforcing standards for interoperability. It could capitalise on the collective strengths of local authorities to meet specific community needs and do what they are there to do - deliver public services for their local communities. It would leverage the scalability of centralised resources and the agility of local innovation, offering a pathway to digital that is both sustainable and responsive to the diverse needs of users.

And of course, no-one has yet properly worked on bringing the local government tech suppliers together to establish standards for the common good. GDS has great form for leveraging the private sector - imagine the impact of a LG-GDS that was similarly inclusive of the established private sector tech firms, helped and supported them, and made them feel included.

So what do we do? We can take a lego-brick approach, create alliances with platform providers and set alignment with common standards and open APIs - which need our smart thinking, ideation, innovation and investment. We can get on with the business of creating standards-based digital services that can be re-used by any council, leveraging services already in play like Gov.uk Notify, Gov.uk Pay and the other stuff councils can and should just plug-and-play consume from their digital platforms, leaving hard integrations and back-office spaghetti apps to those providers.

Strategic relationships with platform providers should be designed to empower local authorities, not to lock them into proprietary ecosystems - and engage with tech providers collaboratively, rather than adversarially, on a shared mission to improve public services.

Best of both worlds

In summary, the potential of open source - flexibility, community-driven innovation - should be used as a complementary force to standard infrastructure, leveraged by digital teams who are solutions people, not developers writing and wrangling code.

In this way, an LG-GDS could leverage the best of both worlds - strategic direction, standardisation, and commoditisation from a centralised approach, alongside the innovation and adaptability of open source communities and emerging open tech, especially when AWS, Google, Meta and Microsoft are all releasing open models right now.

A lego-brick approach means those services can be re-configured, re-used, or swapped out in any context or local setting – instead of creating a mountain of tech debt that can’t be paid down without constantly rebuilding the same thing, over and over and over, while councils make glacial progress with the things that matter.