After months and years in development, your new UK Government policy is ready to implement. Alongside new laws, guidelines or standards, you’re going to deliver a brand new digital tool, portal or one-stop shop. Of course, your Department’s tech teams are all tied up elsewhere, so you’re going out to market to find a supplier to build your new digital service.

You don’t know it yet, but your policy implementation has just entered its riskiest phase.

When a Government IT project goes wrong, the problem often starts way back at the beginning – with the procurement. Lots of Government teams make mistakes in designing their procurements for digital services, end up contracting a supplier who isn’t fit for the job – a lemon – and have a terrible time as a result.

The impact of a flawed procurement can’t be understated. It has seriously damaged the success of numerous Government policies. When projects have missed their delivery deadlines or produced a terrible, buggy service, the impact of good policy is undermined and the Government ends up embarrassed in the press. Not only that, but managing a digital service project that is in dire straits causes considerable stress, distraction and low morale for staff working on these policies, and often drives them away to a different role.

So how do you buy the best team for your service and your policy? What are the common assessment mistakes that undermine quality and the red flags that drive away good suppliers? How do you design your procurement to dodge the lemons?

In this post, we’ll share what we’ve learned about what a good procurement looks like from our decade (and counting!) of delivering digital services for Government, and explain how to identify and buy quality in a digital service supplier.

Ask questions that accurately assess the hardest things to deliver

Many public sector tenders fall into a hidden trap: they focus on maximising value-for-money, when what you as a policy team need is lowest risk (i.e. highest quality) that fits within your budget. Applying a quality/price weighting relies on you actually being able to evaluate quality well – if you can’t, price will dominate in the overall evaluation and you’ll end up with a cheaper lemon. Not only that, but high‑quality suppliers will walk away if they sense the process can’t distinguish their effort from a copy‑and‑paste bid.

Procurements often lean on generic questions that every supplier can answer well. “Do you practice agile methods?” “Can you provide user research?” We’ve even seen procurements where the majority of quality points are awarded for just answering “yes” with no evidence to a list of requirements!

These are table stakes questions – almost every credible bidder will answer positively and most will have stock answers that they will just copy and paste. Questions like these that are too easy (or, indeed, too hard) compress the quality assessment – the weakest supplier looks almost as strong as the best.

Instead, target the hardest parts of your service. What are the real challenges in your project? Is it managing sensitive data, integrating legacy systems, or designing for complex user journeys? Shape your evaluation around those, and ask for evidence that suppliers have these make-or-break behaviours or skills.

Beware of trying to procure too much at once and spreading your quality assessment too thin. If your tender tries to cover a multidisciplinary research function, live service support, and cloud hosting all in one go, you will struggle to properly test for excellence in any one area. You may also end up with a supplier who can do many things to an average degree but nothing well. Scope each procurement to a coherent set of behaviours and skills, and ensure you have at least 8 to 10 questions to assess them well.

Picture your perfect supplier whilst scoring

Now that you have great questions, you also need to use the full range of quality scores. If a response is weak or generic, mark it low. If it’s truly exceptional, reward it. Being reluctant to score harshly or failing to imagine what excellent looks like blurs the distinction between suppliers. Zero is a valid mark – use it when you need to!

An excellent supplier will have put effort into their response because they care about it – they will have worked on the problem as part of the bid, producing artefacts like assumptions maps or prototypes that accelerate you. Good answers will be specific to your needs and will have thought directly about your problem – you should be prepared to penalise obvious copy and paste answers that don’t show thought into your problem.

Choose realistic budgets and timelines

If the pressure is on to get your new digital service live, it’s tempting to pick timelines and prices based on deadlines and available budget. But picking unachievable parameters will result in more challenging contract issues down the line when you’re invested in suppliers that can't deliver. Are you setting yourself up for a delivery that actually works, or one that falters under its own constraints?

Unrealistic budgets are a common trap. Set the number too low and some suppliers will still bid, and then either fail to deliver and demand an extension which you’ll be hard-pressed to say no to, or hand you a cut‑down service that doesn’t let your policy succeed. The ideal budget matches the real effort required, informed by early engagement with your digital and commercial colleagues and the market.

The same holds true for timelines. A short engagement might be theoretically possible, but what happens when dependencies on your side aren’t ready or internal approvals drag? An issue that we commonly see people forget is that getting access to data from other teams takes much longer than they expect – 6 to 12 weeks is common. If the process only starts when the contract is signed, your supplier could be twiddling their thumbs for weeks and your project is fatally behind from day one. Ideally, you need to identify the dependencies and start working on them before your procurement even begins.

Once again, your budgets and timelines are a credibility test. Realistic plans signal to quality suppliers that you understand your own project. Unrealistic ones do the opposite and will push the best providers to walk away from impending disaster.

Cover all the skills and activities you need

Teams often buy services without fully understanding their organisation’s requirements. For example, it’s a misstep to procure cloud hosting without knowing what assurance standards will apply (e.g. does data need to stay in the UK, and what cyber assurance will it need?). In the worst case you won’t find out that you didn’t procure something necessary until your service can’t go live by the deadline – leaving you without any time to react.

Offboarding matters as much as going live in the first place. What happens to your live service at the end of the contract? Who will take over maintenance and support, and if you’re procuring the hosting too, have you built in enough time to migrate the service? Will you be able to get the data out of your supplier, and where will you put it? Standard framework terms rarely cover offboarding in detail, and a lapse here creates risk that surfaces when it’s too late to fix.

Suppliers under pressure on price will reduce team size first, and assessing whether teams have critical skills is often overlooked. Does your supplier’s proposed team know enough about data governance to establish sharing agreements and work proactively with Data Protection? Can they develop data matching when no shared identifiers exist in your datasets, and do they also have the BA capability to build the human-driven parts of that process? Even asking about fluency with SQL can reveal whether a team has real data expertise or just enough to get by. So, have a rough idea of the roles, size and skills you need and make sure you assess them – again informed by the most challenging parts of your service to deliver.

Focus on outcomes, not specific solutions

When you write your digital service tender, your instinct might be to specify as much of the solution as you can. Procurements we see are frequently over-confident about what the right technologies, tools, or certifications are. It seems defensive – after all, a more tightly specified procurement makes it more likely that you get what you want, and helps select a supplier who can do exactly that, right?

You might have very specific ideas about how your service should work today, but if you’re early on in the delivery of your service (i.e. in Discovery or Alpha), what you learn during delivery often means your ideas will need to change. When we work on digital services in early phases, we normally always find that the best path to success is not quite what the customer expected initially (and they’re thrilled that we found that out early!).

Overly prescriptive requirements select for teams that can implement instructions but won’t learn and won’t iterate the vision. This can cause a failed Service Assessment or (worse!) lead to a poor quality service that users complain about publicly. They also filter out suppliers with expertise that might even show you a faster, safer or cheaper approach. Do you really need a Microsoft Gold Partner, or do you need a mature team that can adapt to solve your problem?

Instead, be honest about what you don’t yet know and focus on the outcomes you need. Define the problem to be solved, the user need to be met, and the policy goal to be achieved, and select suppliers that will work with you to develop the solution as you learn. Use quality evaluation questions to assess problem solving, critical thinking, and collaborative working, aiming to select a supplier who can solve complex problems by adapting to new findings.

Be general about required domain experience

Another common trap lies in asking for very specific domain experience. Working with domain experts is absolutely necessary, but we often see procurements that ask for prior expertise with specific policies, datasets or legacy systems. Whilst this will rapidly cut down the pool of potential suppliers, it will also exclude suppliers with broader experience in your domain that may be better overall. Do you really need a supplier with niche knowledge of a certain genomic sequencing dataset, or would a supplier with an excellent track record of delivering genomic data services successfully be less risky overall?

On the other side of the spectrum, we also see that tender questions asking for “experience in similar projects” often backfire. Which aspects need to be similar: policy area, user type, scale, or technology stack? There are few government services that are truly alike. Ambiguity pushes suppliers to stretch their experience to fit, and implicitly encouraging that frustrates your efforts to assess supplier quality well.

Instead, find a middle ground: identify the core similarities that really matter and select domain experts who can adapt their experience to your niche. A strong supplier will demonstrate capability to learn quickly, adapt, and integrate with your data landscape. Those qualities allow a supplier to deliver a resilient, user‑centred service even in unfamiliar territory.

High quality tenders attract high quality suppliers

Picture yourself at the final hurdle. The contract is signed, the supplier is on board, and your service is about to begin delivery. This is the moment when every choice in your procurement process either pays off or comes back to haunt you – did you pick a lemon? Have you truly assessed the toughest aspects of your project? Have you created the space for suppliers to demonstrate quality over price? And are your timelines and budgets realistic, not just hopeful?

If you hesitate on any of these, it’s time to pause and seek advice. Skipping early engagement with Digital and Assurance colleagues feels faster, but it rarely is. These teams can spot risks in your scope, evaluation, or timelines before they become unmanageable. Their assurance isn’t a box‑ticking exercise – it’s a safety net.

Here’s the secret: successful digital procurement isn’t about speed, it’s about certainty. A well‑designed tender attracts the right suppliers, tests for the hardest parts of delivery, and builds credibility. A rushed or poorly‑scoped procurement all but guarantees stress, reputational risk, and wasted effort.

So get it right the first time: involve your digital experts, do pre-market engagement, design tough and fair evaluations, and respect the realities of time and cost. Your future self – and your policy – will thank you.

Authors

Simon Worthington

Simon is an architect and engineer working to build better interfaces between humans and data. Sometimes that’s by leading new products developing cutting-edge data infrastructure and sometimes that’s by researching new patterns and models with real data users.