The last mile: why data still isn't reaching council decisions

Ask anyone in a council whether they have enough data and you will rarely hear no. Most have finance systems, case management systems, performance dashboards, programme trackers and, almost without exception, a forest of spreadsheets holding the whole thing together. What they are short of is a dependable way to get the right part of that data in front of the right person before a decision is made.

There has been plenty of good writing on the data side of this. But I want to look at the other end of the pipe. The cabinet paper, the programme board, the corporate leadership team meeting where somebody has to say yes, no or not yet. The research published over the last two years points to that last mile as the place where things most often go wrong. It is also where the right tools can make a difference quickly, without waiting for a five-year data strategy to land.

What the research actually says

The evidence now comes from several directions: the sector's own surveys, central government, auditors and inspectors. Read side by side, it tells one consistent story. Councils hold more data than ever, but too little of it reaches decision-makers in a form they can trust, understand and act on.

Council data leaders say they are early in the journey

Between October and December 2024 the LGA surveyed heads of data in every English council and combined authority, and published the results in April 2025. Ninety-six organisations replied. The LGA is careful to call it a snapshot rather than a fully representative picture, but the pattern is hard to ignore.

• Only 14% placed their organisation in the top two of five maturity levels. Most sat at level 2 or 3.
• Three quarters said poor data quality was a challenge to a great or moderate extent.
• 62% said staff had too little data literacy to do their jobs well, and the question specifically included using insight to support decisions.
• Every respondent still relied on Excel, and 70% said they were only at the start of their AI journey.
• 44% found it hard to evidence the impact of investing in data at all.

One line in the free-text answers deserves more attention than it gets. Several heads of data said their analysts lose heart when good work never changes anything, either because the organisation is not ready to use it or because decisionmakers push back. That is the last-mile problem in a sentence: the analysis exists, and then it stops short of the decision.

Government's own review rated data as the weakest area

The State of Digital Government Review, published by DSIT in January 2025, drew on a digital maturity survey of 76 local authorities. Using and managing data came out as the lowest-scoring area of all. The same review found that digital and data roles make up just 2 per cent of local government headcount, half the 4 per cent benchmark.

Across the wider public sector, only 27 per cent of leaders surveyed for the review felt their data infrastructure gave them a comprehensive view of operations. Seventy per cent said their data landscape did not provide a unified source of truth.

Welsh auditors found leaders cannot see outcomes

Audit Wales reviewed all 22 Welsh councils and asked a simple question. Does the performance information that reaches senior leaders help them understand what residents experience, and what the council's activity is actually achieving? Its national summary in July 2024 found that, apart from a few councils, leaders were given limited information on either.

In Bridgend, for example, auditors concluded that this limited leaders' ability to understand the impact of the council's own services and policies. It is a Welsh review, but anyone who has sat through a quarterly performance report at an English cabinet will recognise the description.

When councils fail, poor information sits near the centre

In March 2026 MHCLG published its Lessons learnt from best value interventions, the first structured study of all 24 councils that had been through intervention since 2020. The number of councils in statutory intervention rose from one in 2020 to nine in each of 2024 and 2025. Governance problems featured in 18 of the 24.

The report is blunt about data. Both the fieldwork and the literature suggest councils do not make enough use of the performance data they already hold. Participants traced this partly to the 2014 abolition of the Audit Commission, which thinned out corporate data and analysis teams and left weak performance monitoring behind.

It also notes that audit delays since Covid have left some councils making decisions on incomplete or inaccurate financial information. One written submission went further, pointing to frequent oversight gaps around transformation programmes and a broken "golden thread" between strategy, planning and delivery.

Inspectors keep finding the same pattern

The Spelthorne best value inspection, published in March 2025, found that poor, late and incomplete reporting had "severely undermined informed decision making". Inspectors said reports to members often lacked rigour, the risk registers were not fit for purpose, and they could not be confident that every figure the council had given them was correct.

Birmingham's Oracle programme is the better-known example. In February 2025 its external auditor, Grant Thornton, published a public interest report on the failed implementation. Officers knew about the problems through 2022 and into 2023, yet the auditors could not find a cabinet or committee report that set them out clearly.

Risks sat in the detail of reports rather than the key messages, and reports to members were judged to acknowledge issues while minimising them. Members said they lacked the capacity to scrutinise properly and relied on officers and delivery partners. The total cost has since moved from an early business case of £19 million to around £131 million. The council says it has since strengthened its governance. I mention it because it shows so clearly how information can exist inside an organisation and still never reach the people making the decision.

What the gap looks like in practice

Put those findings together and the problem is rarely that the data does not exist, it is not being put together in a way that is accessible. From what we see across our own clients, it tends to go wrong in four ways.

• It arrives late - In a lot of councils the monthly or quarterly report is still put together by hand, from updates chased by email and pasted into a template. By the time it reaches a board it describes a position that is six weeks old, and decisions get made on last quarter's picture.
• It arrives without meaning - A page of red, amber and green tells a councillor what somebody thinks, not why they think it. A project that has been amber for three quarters looks exactly the same on paper as one that turned amber yesterday. This is what Audit Wales and CfGS were describing: plenty of reporting, very little that explains outcomes or trajectory.
• It arrives softened - Status is usually set by the person closest to the work, who is also the person with the most reason to hope it comes good. Nobody is being dishonest, but optimism compounds as a report travels up a hierarchy.
• It cannot be questioned - Even when the right information is somewhere in the pack, decisionmakers rarely have the time or the tools to interrogate it. If the only way to test a number is to email an officer and wait a week, most people will not.

None of these is a data quality problem in the usual sense. A council could have spotless data and still fall down on all four. That is why fixing the foundations, important as it is, will not close the gap on its own.

Why it shows up first in programmes and projects

A large share of a council's most expensive and least reversible decisions run through programmes and projects. Transformation and savings plans, capital and regeneration schemes, system replacements, SEND and children's services improvement. This is where money is committed up front, benefits arrive later (if they arrive at all), and where MHCLG's study found oversight gaps cropping up again and again.

Programme data is also unusually scattered. The plan lives in one tool, the risk log in another, the money in the finance system and the benefits in a spreadsheet that belongs to whoever set it up. Pulling that together for a board is exactly the manual, late, softened process described above.

The stakes are rising, too. Medium-term financial plans now lean heavily on savings programmes delivering on time. When one slips it stops being a project problem and becomes a budget problem, usually discovered at the point it is hardest to fix.

Reorganisation and uncertainty raise the stakes

Local government reorganisation adds another layer of complexity. Following July’s announcements, the government paused the wider programme for review in September 2026 and withdrew its decisions for Essex, Hampshire, Norfolk and Suffolk. Surrey’s reorganisation remains unaffected.

For councils affected by the review, the future structure and timetable remain uncertain. But existing services, savings programmes and capital projects still need to be managed, with leaders able to see costs, risks, dependencies and expected outcomes clearly.

Where reorganisation proceeds, bringing together information from councils with different systems and reporting practices will be a substantial task. Agreeing common delivery data and status definitions can support current oversight while making any future transition easier.

Closing the gap: what actually changes a decision

None of what follows needs a new data warehouse. It is about the route from data to decision, and most of it can start this quarter.

1. Start with the decision, not the dataset - List the decisions your leadership team and programme boards will take in the next twelve months, and what each one genuinely needs to know. Most councils that do this find they report a great deal nobody uses, and miss the two or three things that would change the answer.
2. Keep delivery data in one place - If status, risks, issues, costs, benefits and dependencies live in a single system, board reports become a view of that record rather than something rebuilt every month. It also ends the two-teams-two-numbers argument before it starts.
3. Make status something you can show your working for - A RAG rating should trace back to evidence: milestones missed, spend against budget, risk exposure, forecast benefits. Show how long something has held that status and which way it is heading and put outcomes next to milestones.
4. Shorten the reporting cycle - When reports are generated from live data, the PMO's time goes into analysis and challenge instead of chasing updates and formatting slides. A board that sees this week's position makes different decisions from one that sees last quarter’s.
5. Let decisionmakers ask their own questions - A director or a scrutiny chair should be able to ask which savings schemes are least likely to land this year, and why, and get an answer the same day without waiting for an officer to build one.
6. Make bad news cheap to raise - Culture matters most here, but tools help. If early warnings surface automatically from the data, nobody has to stake their reputation on being the first to say a project is in trouble.

For councils preparing for possible reorganisation, agreeing common delivery data and status definitions can strengthen oversight now and support any future transition.

Where Verto Intelligence fits

We built Verto Intelligence to work on that last mile. It sits inside Verto, on the live programme, project and portfolio data councils already keep there, rather than asking anyone to export data into a separate AI tool. Our principle is simple: bring the AI to the data, not the data to the AI.

Data Insights: a second opinion on status

Data Insights puts a small icon next to status fields. Click it and Verto Intelligence reviews the project's scope, costs, benefits, milestones, risks and issues, then recommends a RAG status with a written explanation.

The project manager can accept it, adjust it afterwards, or dismiss it and set their own. Anything generated this way is flagged in the change history, so a board can see where a status came from. When the person and the evidence disagree, that is exactly the conversation a programme board should be having.

Verto Intelligence Partners: specialists you can question

Verto Intelligence Partners, or VIPs, are 18 AI specialists, each focused on one discipline: risk, finance, benefits, IT and so on. A coordinator gathers their views and summarises them. You ask a question in plain English, and they will explore the data, suggest changes such as a revised status, and add risks, issues, costs or milestones when asked.

Two details matter for decision-making. First, each partner can be given a personality, so you can choose a deliberately pessimistic risk manager or a questioning benefits manager. That is a useful voice to have in the room when everyone else would like the project to be green. Second, partners can look across a whole programme or portfolio, which makes questions like "which of our savings schemes is least likely to land this year, and why?" answerable without anyone commissioning a report.

Before switching them on, we ask each organisation for its context: strategic objectives, constraints, background. That way the answers reflect the council's priorities rather than generic project theory. For the many accidental project managers in local government, it is also a patient expert to learn from.

Meeting councils where their data actually is

Every council in the LGA's survey still relied on Excel, so we started there. AI-assisted import reads a spreadsheet of risks, issues, costs, savings, benefits, lessons learned or milestones and maps it to the right fields in Verto, with a review step before anything is saved. Verto also works alongside Microsoft 365, Teams and Power BI, so insight can reach people in the tools they already use.

Built for public sector governance

All of this runs within Verto's UK-hosted environment, which holds ISO 27001 and Cyber Essentials certification. [Internal technical confirmation required before publication: the partners have no internet access; cannot send data outside the council's Verto site; and only see records the person asking has permission to see.]

We are also straightforward about what AI should and should not do here. Verto Intelligence suggests; people decide. Our own guidance tells users to review every recommendation before accepting it, because the same question can produce a slightly different answer each time. Keeping a person accountable for the decision, with a clear record of what the AI suggested, is part of good governance rather than a limitation of it.

A better question to ask

Every council should be working on its data foundations, and many are. But there is a quicker test of whether data is actually shaping decisions.

Take the last significant programme decision your cabinet or leadership team made. What did the paper rely on, and how old was that information? Who set the status, and could anyone in the room have challenged it there and then?

If those answers are uncomfortable, the gap is probably not in your data warehouse. It is in the last mile, and that can be fixed a lot faster than most people expect.

If you would like to see what this looks like on your own programmes, our maturity assessment is a good place to start, or you can book a session to see Verto Intelligence working on real project data.