GoC’s Shared Services Represents Only Part of the Solution

The Government of Canada announced an initiative to centralize services across the government, to develop consistency across departments, and most importantly cut costs by reducing redundant systems and contracts.

On paper, it makes sense. To improve efficiency and lower costs by consolidating services centrally and reducing contracts. Solid argument, seems very reasonable.

Sadly and unsurprisingly, in practice, there is no direct correlation between simply centralising services and realising cost benefits. Despite many attempts before to centralise services, the result hasn’t always been savings in cost, but instead the opposite.

The problem isn’t with the lack of shared services / centralised services, but that the very framework of shared services is built on a broken procurement model that:

• rewards and reinforces unintegrated systems,
• prevents non status-quo systems, and FOSS / Open-Source options;
• is built on reinforcing sustainability of the framework for increasingly greater financial rewards.

This procurement model in government may have worked when times were simpler. Except at that time finding valid & qualifiable (security checked) vendors was harder, and systems were less complex – nothing changed and these systems would last for years. Fast-forward, and you have the mess government has now:

• incapability to meet the varied technical needs of a 250’000-strong federal workforce (what? You want to use Firefox or Chrome browser?)
• incapability to procure modern software or hardware for specific or unique projects (forget getting any Apple Macbook with Adobe Illustrator);
• a fragile technical infrastructure strung together by outdated systems, protected by blocking sites and even full blocking public servants’ access to the Internet (because threats are looming);
• procurement of costly proprietary systems that are adapted or customised to other outdated proprietary infrastructure (often Windows Servers);

To make it worse, combine this with a complex society with ever-changing needs and increasing demands on government to provide professional services on par with the private sector, and you’ve got an urgent mess.

This mess goes further. The big actors in procurement to the Government of Canada are CGI and IBM, two huge giants in the business with first dibs and first call to big contracts, with keys to the infrastructure gates to deem what other software and vendors can work on the Government’s custom-built infrastructure. As Government works to cut costs and reduce a supposed “number” of concurrent systems, these big vendors maintain an inside track of access, expertise and even input (or “influence”) on the selection of systems. How systems get approved to go on this infrastructure is mired in kafka-esque bureaucracy reinforced by techies promoted through the ranks of playing it safe (remember, “Nobody ever gets fired for choosing Microsoft technology”), rationalised by subjective risk matrices matched with supporting argumentation for the strength of status quo. Status quo is fine and good, so long as things don’t change. And things change in the tech world. Lots.

So government departments, divided and conquered by Ottawa’s big vendors, buy duplicate systems they can’t maintain, that are closed to support an update cycle paid for by successive licensing, at inflated rates to work on a disparate infrastructure. Centralising the services won’t resolve this; instead it just may make it easier for big vendors to collect more money with fewer invoices as they eliminate more smaller vendors who were offering alternative solutions for custom problems.

BUT the tide is changing. Maybe. The elephant outside the room is demanding more from government. All the while new hires and resistant public servants continue to voice objection to working with inferior disconnected PCs with inferior unsocial software on 90s-era infrastructure.

More silver-lining may lie in that the infrastructure gatekeepers (they’re on the government side) seem largely agnostic when it comes to types of systems. I can relate that it’s unnerving for a public servant to endorse a FOSS (Free Open Source Software) tool if there’s no one on the other end of the line to fix the system that a thousand employees may rely on. Big vendors are good at providing guarantees to getting the job done according “to spec”, a sure-sell in a public service built on risk-avoidance. Matched with a good procurement specialist who can hold them to their penalties when they under-deliver, and cost savings are realised. Unfortunately, the sad reality is considering the level of turn-over in procurement and lack of IT specialists involved in the procurement process (for some reason, they seem to be excluded), the big vendors win out like a game of Risk when your opponent is not only ill aware of the board’s pieces, but they’re playing a different player every few turns.

The Government of Canada wants to increase efficiency of systems and lower costs. The existing model to procure systems is broken and costly. With 44 departments and agencies using very complex and costly systems, saying centralising the services will resolve the issue is oversimplistic and attacking the symptom, not the cause. To centralise systems is analogous to fixing a complicated game of Football by having 44 players play on a smaller field, when it’s the rules that need to be changed. The government ought look at the procurement process, with a criteria to realise efficiencies and cost savings, with centralisation acceptable only as an option among any multi-tiered approach for supporting a range of deparments with varying needs and challenges (hey, for some departments it works, others, it doesn’t).

Procurement is the process of the acquisition of goods and services according to agreed terms for money. If the government is having a problem with the goods and services, and wants to save money, then it’s time to change the process. Not put all your eggs in the same broken, costly basket.

Doug Bastien