New Oxford research reveals the potential impact of NPM 2018 on food and drink categories, and what it could mean for businesses.
New research from the University of Oxford suggests that applying the updated Nutrient Profiling Model (NPM 2018) to existing HFSS restrictions could increase the number of products in scope by almost a quarter. Soft drinks, breakfast cereals and yoghurts are among the categories most affected.
The researchers assessed 15,514 products across 13 categories covered by existing restrictions, using product information collected from the websites of Tesco, Sainsbury's, Asda and Ocado, including M&S products.
Under NPM 2018, 72% of products were classified as high in fat, salt or sugar (HFSS), compared with 58% under the current NPM 2004/05.
The study was funded by the National Institute for Health and Care Research (NIHR), with research questions co-developed with the Office for Health Improvement and Disparities (OHID). Although the University of Oxford conducted the analysis independently, OHID's involvement suggests the findings could be an important contribution to government's assessment of the proposed changes.
The timing is significant. The consultation on applying NPM 2018 to existing advertising and promotions legislation in England closed on 17 June 2026. A 12-month implementation period was proposed, but the government response and final decisions are still awaited.
Where would the biggest changes be?
Of the products analysed, 84% received the same classification under both models. Of the 16% that changed, 93% moved from passing under NPM 2004/05 to failing under NPM 2018.
That means almost a quarter more products will potentially be subject to restrictions. Importantly, this is based on the number of products (SKU count), not the proportion of sales.
However, the impact varied considerably by category.
The proportion of products passing the model fell particularly sharply for:
Soft drinks: 66% to 6%
Breakfast cereals: 79% to 39%
Yoghurts: 78% to 36%
Pizzas: 51% to 29%
Morning goods: 32% to 13%
Ice cream and desserts also saw substantial reductions. In contrast, confectionery, cakes and biscuits saw relatively little change, as most products already failed the existing model. Ready meals and savoury snacks saw more modest reductions, while processed potatoes remained unchanged.
The main driver was the move from total sugars to free sugars, alongside tighter scoring thresholds. Increased fibre points offset some of the higher sugar scores in certain categories.
For businesses, the message is clear: the impact of NPM 2018 will depend heavily on your product portfolio. Category averages are useful for understanding the direction of change, but individual product assessments will be essential.
Free sugars: the unresolved challenge
One of the biggest practical challenges is that free sugars are not declared on nutrition labels, as they can’t be measured in a lab, making consistent calculation across products and businesses difficult.
The Oxford researchers estimated free sugars using ingredient lists and estimated ingredient quantities. Broadly, rather than adding up sources of free sugars, they estimated the sugars contributed by ingredients not considered sources of free sugars, such as milk and whole fruit, and subtracted these from the total sugars declared on pack.
This provided a valuable way to model thousands of products using publicly available information, for academic purposes. Sensitivity analyses suggested the overall findings were relatively robust, although some categories and individual products close to classification thresholds were more sensitive to the assumptions used.
Importantly, this was a research methodology, not a validated regulatory calculation method.
Manufacturers may have access to detailed formulation data, retailers generally depend on information from suppliers, and researchers often rely on publicly available ingredient lists and nutrition declarations. Different approaches and assumptions could produce different free sugars estimates, potentially changing individual product classifications.
As IGD's previous work with industry has highlighted, a consistent, scalable and auditable methodology will be essential to give businesses and regulators confidence in how the model is applied.
What about reformulation?
The study also considers the potential for reformulation. Overall, the proportion of failing products sitting close to the pass/fail threshold increases only modestly, from 10% under NPM 2004/05 to 13% under NPM 2018. The authors suggest a similar share of products could potentially be reformulated to pass.
While this could help businesses continue to prioritise categories for NPM and reformulation, being close to the threshold on paper does not necessarily mean a product can be reformulated easily or commercially. The scope to reduce free sugar, for example, will depend on the product, its role in the recipe and the implications for taste, texture, and shelf life. It will also depend on customer expectations – a strawberry yogurt could have less free sugar if less strawberry puree is added, but the flavour profile will change. Businesses will need to assess opportunities at product level rather than assume that category-level findings translate into straightforward reformulation options.
What does this mean for Mandatory Reporting?
The findings are also relevant to the wider discussion around Mandatory Health Reporting, but it is important to distinguish between individual product classifications and sales-weighted reporting.
The Oxford study assessed the proportion of products passing or failing each model. It did not account for how much of each product was sold, so the results cannot tell us how a retailer's overall healthiness of sales would change.
We are still awaiting the government's consultation on Mandatory Reporting, including confirmation of which nutrient profiling model and metrics may be used. How free sugars would be calculated consistently at scale also remains an important question.
The research therefore provides useful evidence on the potential implications of NPM 2018 but does not resolve all the practical requirements for HFSS legislation or Mandatory Reporting.
A changing picture across the UK
Meanwhile, businesses are navigating different approaches to HFSS restrictions across the UK.
In England, placement restrictions have applied since October 2022, with volume price promotion restrictions introduced in October 2025. Wales introduced its promotion and placement regulations in March 2026, while Scotland's regulations came into force on 1 October 2026, using NPM 2004/05.
Northern Ireland's consultation on proposed restrictions closed on 21 October 2026, with decisions on the final approach awaited.
If different nutrient profiling models are adopted across nations, or used for different policy purposes, businesses operating UK-wide could face multiple definitions of an HFSS product. This would add complexity to product assessments, data systems, reformulation planning and compliance.
The Health and Social Care Committee's recent report, The obesity moonshot: making obesity a political priority, reinforces the need for urgent, coordinated action on obesity. Consistency across policy will be important if that ambition is to translate into meaningful change.
What should businesses do now?
Although policy decisions are still awaited, the research provides a useful starting point for businesses to understand their potential exposure to NPM 2018. Practical steps include:
Assess your portfolio. Identify which products may change classification under NPM 2018, using your own product and supplier data
If you are a retailer – look at Appendix 5 in the paper which documents the ‘Number and proportion of products passing and failing NPM 2004/5 and NPM 2018 by retailer and in-scope category’
Understand your free sugars data. Identify information gaps, sources of free sugars and where calculation assumptions may be needed. The study's Appendix 1 provides a useful reference list of ingredients used in its estimation approach, alongside the OHID technical guidance document on applying NPM 2018.Prioritise borderline products: explore where reformulation may be technically and commercially viable, recognising that a small change in score does not necessarily mean a simple product change.
Plan for different requirements: consider how product data and systems would cope with different models being used across nations or policy areas.
Share practical evidence: feed the realities of calculating free sugars at scale into industry and government discussions, helping to develop a consistent and workable approach.
Read the full research paper: to see how the paper also compared NPM 2018 with different front of pack labelling schemes including UK multiple traffic light labels (MTL), Nutri-Score, and the Chilean Warning Labels.
Read the full study: Sarda, Scarborough and Bandy (2026), University of Oxford. DOI: 10.5287/ora-mv5w9mmx8.