This time it's personal: the rise of dynamic, personalised pricing and what it means for inflation

The rise of algorithmic pricing

Technology is enabling this shift. Digitalisation has radically reduced what economists call menu costs – the expense of changing listed prices – so called historically because of the costs associated with having new menus printed with updated prices. Digital pricing allows firms to change prices frequently at negligible cost. Online travel and hotel platforms, once simply passive listings sites, now dynamically alter their prices as demand ebbs and flows. For instance, the proportion of hotel room rates that change at least once per month has risen from around 15% in 2005 to roughly 80% today (Chart 1).

Chart 1: Share of CPI price observations for UK hotels changing month to month

Footnotes

• Notes: Accommodation services are mostly driven by hotel prices. The orange line is a representative comparator series constructed by removing sectors with prevalence of dynamic pricing from CPI services. The aqua line has breaks in 2020 and 2021 as hotel prices could not be collected during the Covid pandemic lockdown periods. Both series are adjusted to exclude the price changes associated with the 2008, 2010 and 2011 changes to VAT.
• Sources: Office for National Statistics (ONS) and Bank calculations.

Technology has also allowed businesses to collect, process and store more data on consumers and their shopping preferences. Meanwhile, machine learning models are enabling more finely tuned price setting using this data. Instead of relying on simple rules like peak and off-peak fares, firms increasingly deploy predictive models that infer demand curves and monitor competitors to optimise their own prices.

In many cases, consumers now see price promotions unique to them or to people with characteristics or shopping habits similar to them: loyalty card discounts, customised bundles or subtle nudges to steer them toward higher margin products. The result is different customers paying different prices for the same or very similar goods and services. Through these practices firms are seeking to move towards the textbook view of ‘perfect price discrimination’ – charging as close to the maximum price a consumer is willing to pay for a good or service.

How much dynamic and personalised pricing is going on?

The sophistication of such technology-enabled pricing varies across the economy. It’s difficult to know precisely how adoption and use vary across sectors because, understandably, firms keep their pricing strategies a closely guarded secret. But we can get a sense from the Bank’s Decision Maker Panel (DMP) Survey of businesses (Chart 2). Over the three months to January 2026 this survey asked over 1,600 firms if they used:

Rules-based pricing – prices that follow simple pre-set rules or vary by time of day, week or season (for example, peak and off-peak train tickets) or automatic adjustments when stock levels change.

Market-responsive pricing – where firms use algorithms, including artificial intelligence (AI), to adjust prices in response to data on demand, capacity or competitors’ prices.

Personalised pricing – prices that differ across customers based on characteristics or behaviour. This could be simple strategies such as loyalty schemes or more sophisticated algorithms applied to data on individual or group shopping habits.

Different businesses may be using some, all or none of these practices in setting their prices.

Chart 2: Usage of different pricing strategies as reported by firms in consumer-facing sectors

Footnotes

• Notes: Bars show current adoption of different pricing strategies. Diamonds represent expected adoption of different pricing strategies in the next 12 months. Percentages may sum to more than 100% in a sector as businesses may be using more than one strategy. Other services include sectors such as education, professional organisations and other personal service activities businesses.
• Sources: DMP Survey (November 2025, December 2025 and January 2026 waves).

Adoption varies. All sectors use these strategies to different degrees and all sectors are proposing to increase their use over the coming year. The largest increases are planned in the use of market-responsive pricing tools, from 21% of firms currently to 31% in a year’s time as more firms utilise data about demand conditions and their competitors to determine their prices. This is consistent with what we hear through the Bank of England’s Agents’ network of business contacts around the UK. We are also seeing some sectors experimenting with technology that could enable dynamic pricing in the future, such as electronic shelf labels in supermarkets, which are already widespread in Europe.

Simple rule-based dynamic pricing is widely used in recreational services, for example, cinemas charging higher prices during peak demand times like weekends but it is less prevalent in other sectors. Anecdotal evidence suggests this method is becoming outdated and replaced by market-responsive dynamic pricing using algorithms; some theme parks now use algorithms that use actual capacity data (how many tickets have been sold in real time) rather than weekday/weekend pricing.

Personalised pricing is widely adopted across all sectors. Largely, these are still simple forms of personalised pricing, for example using loyalty cards and online customer accounts, and are widespread in recreational services like gym memberships.

Adoption in the UK may be limited more by reputational considerations than by technological ones. Businesses may worry about the potential consumer backlash from changing prices in a way that is opaque and thought unfair. International studies show that consumers in the UK are more likely than consumers in other countries to consider dynamic pricing unfair.^{footnote [3]} Businesses may also be wary of regulatory actions.^{footnote [4]}

Challenging inflation measurement

A world of more fluid and bespoke prices could make life harder for statisticians, while the power of data analysis may make it easier. The consumer prices index (CPI) currently relies on collecting a representative sample of prices for a standardised basket once a month. That works well when prices mostly move slowly and uniformly. But when prices shift continually –and differently for each shopper – the idea of a ‘representative’ price becomes strained.

Price volatility is one challenge. Hotel and airfare prices, now highly dynamic, increasingly inject noise into the monthly inflation reading. As firms change prices more frequently, month-to month inflation becomes harder to measure and interpret. Yet the greater flexibility also means that shocks may pass through more quickly to prices,^{footnote [5]} meaning some of those jagged movements may actually be informative.

Central banks are responsible for stabilising overall inflation. What matters to them is the overall trend of inflation, rather than any particular monthly movement in the CPI. This often involves looking at measures of inflation that are less volatile and provide a clearer signal of the underlying trend. For example, at the Bank of England, we exclude hotel and airfares prices, where prices are set dynamically, from our CPI ‘core services’ measure, which we use to help monitor underlying inflationary pressures. That is a relatively simple approach. More broadly, in recent years, central banks, including the Bank of England, have expanded their analytical capability to extract signals from many different data sources to help identify the economic trends beneath the surface of the month-on-month volatility that is typical of economic data.^{footnote [6]}

Another challenge is that the rise of personalised pricing splinters the consumer experience and raises the question of what it really means for a price to be ‘representative’. Households already face different inflation rates because they buy different things, in different ways. Inflation rates experienced by different income deciles vary because of their different consumption baskets.^{footnote [7]} Although this has always been the case, these different consumption baskets can become more segmented with personalised pricing. And when prices differ for the same thing, inflation becomes even more personalised – and aggregate measures may no longer reflect households’ experience.

Statistical offices are adapting. In March 2026, the Office for National Statistics (ONS) started to use bulk weekly grocery scanner data.^{footnote [8]} This means using far more prices in the CPI measure, collected over a greater time period, and for the first time capturing loyalty card prices and person-specific discounts from retailers representing around half of the grocery market. This is a small revolution in the measurement of inflation.

Evidence on scanner data shows the impact is very small. Early analysis estimates that supermarket scanner data would have reduced headline CPI inflation by 0.03 percentage points from January 2022 to June 2025. But for specific periods of time, the impact on headline CPI inflation might be larger. For instance, headline inflation would have been about 0.1 percentage points higher during months when food price inflation was particularly elevated. And for individual goods and services the differences could be larger. However, these relatively small overall effects will continue to mask much larger differences between the prices faced by different households purely because they consume different baskets of goods and services. In recognition of this, in 2023 the ONS started publishing quarterly bulletins with Household Cost Indices which provide information on the prices paid by different types of households.

Will dynamic pricing raise inflation?

For now, algorithmic pricing does not appear to be an inflationary menace and is used more to manage capacity than raise prices. Returning to those hotel prices, we can see in Chart 3 that despite the explosion in dynamic pricing since 2005, the cumulative increase in accommodation prices is broadly in line with the prices of less dynamic services.

Chart 3: Cumulative increases in the aggregate price level of hotels since 2005

Footnotes

• Sources: ONS and Bank calculations

Perhaps this should not be a surprise: dynamic pricing and greater price discrimination ought sometimes to lead to prices being higher than normal and at other times lower than normal. As well as potentially enabling firms to charge higher prices to those willing to pay more, more personalised pricing might also enable consumers who had previously been unable to afford certain goods and services to access them.

Moreover, by allowing prices to vary with capacity, dynamic pricing means better use of spare capacity and increased overall economic activity. Dynamic pricing can increase capacity utilisation by better matching supply and demand, lowering costs. Airlines and ride-hailing platforms use surge pricing to deploy capacity when it is needed most. Over time, such efficiencies should increase activity and reduce prices. But for these pricing strategies to benefit consumers, they will need to have fit-for-purpose safeguards to ensure consistency both with good ethical standards and the law. The Competition and Market Authority (CMA) have issued guidance to ensure transparency, with clear legal and reputational implications.^{footnote [9]}

How dynamic and personalised pricing will affect inflation will depend on a whole range of factors, in particular the degree of competition in a market and the availability of information about prices to consumers. As well as how it affects the formation of households’ inflation expectations.

Information and competition

Data advances give firms unprecedented insight into what customers are willing to pay. In theory, this allows them to price-discriminate better and capture more consumer surplus – the benefit consumers get from purchasing a good or service. In markets with limited competition, mark ups could rise, and with them, prices.^{footnote [10]} But where rivalry is fierce, better information on customers can lead firms to undercut competitors, keeping a lid on prices. Moreover, customers also increasingly have access to better information on the firms offering a product. This increases consumers’ price elasticity of demand, limiting firms’ ability to raise prices.^{footnote [11]} AI shopping assistants already help people hunt for bargains, though benefits accrue mostly to digitally savvy consumers who use AI in their search for the lowest prices.

Yet in some unsettling experiments, algorithms have learnt over time to charge above competitive and static prices and drifted into tacit, collusive behaviour after discovering that constant undercutting is bad for profits.^{footnote [12]} Algorithmic collusion is still rare, but regulators are alert.

Inflation expectations

Psychological channels will likely be important. Households notice price rises more than price falls, and form inflation expectations on prices of goods and services they purchase and therefore see often, such as food and fuel.^{footnote [13]} If dynamic pricing becomes more common in those categories, more volatile prices may cause perceptions of inflation to drift upwards – even if average prices do not. The latest results from the Bank’s Inflation Attitudes Survey show that when prices fluctuate because of AI-enabled dynamic pricing, people perceive that prices are higher on average,^{footnote [14]} which may lead to them forming higher inflation expectations in the future.

Chart 4 plots the prevalence of dynamic prices across different items against their importance for forming inflation expectations. It shows dynamic pricing is somewhat limited in sectors which sell items that have the most bearing on forming inflation expectations. However, food is an outlier as it has a high degree of dynamic pricing and a high effect on people’s expectations of inflation. We expect these pricing strategies to be increasingly used across most sectors, including in food, which could lead to an upward ratchet on inflation expectations as food prices become more volatile and consumers notice the increases more than the decreases.

Chart 4: Price change frequency and impact on household inflation expectations by sector

Footnotes

• Note: This chart only shows sectors that influence formation of households’ inflation expectations as in Anesti et al (2025).
• Sources: DMP Survey, ONS and Bank calculations.

It matters not just that central banks understand what drives pricing, but that consumers do too. If consumers understand that price fluctuations reflect demand conditions and that higher prices are short-lived, expectations can remain anchored. We see this in rail travel, where peak/off-peak pricing is well communicated and widely understood. Transparency can reduce price uncertainty and keep expectations anchored when prices fluctuate.

Conclusion

Dynamic and personalised pricing have not so far radically altered the overall inflation process. But they are changing it, and the direction of travel is clear. As data-driven pricing, particularly driven by new technologies such as AI, becomes more prevalent, prices will move faster and vary more across consumers as well as for firms. This creates new opportunities for both. Central banks need to keep track of how this affects consumer behaviour and inflation and continue to improve our tools for extracting information about underlying economic conditions from more volatile, more fragmented, prices. We also need to do more to understand how households and firms experience inflation and what this means for their inflation expectations.

Share your thoughts with us at Bankinsights@bankofengland.co.uk