Technical Appendix
About our survey
An online survey of businesses in digital advertising was fielded in the UK between 18th Feb - 27th February 2026. The sample comprises a total of 258 practitioners in the sector. Public First is a member of the BPC and abides by its rules. For more information please contact the Public First polling team: polling@publicfirst.co.uk
Chapter 1
GVA impacts of digital advertising
To quantify the economic impact of digital advertising in the UK, Public First had to draw on a range of data sources, make reasoned assumptions in places and apply a standard approach to economic impact modelling.
The measure captures economic output (GVA) and jobs supported by digital advertising in the following areas:
- The advertising and marketing sector (e.g. ad agencies);
- Advanced advertising: ad tech, retail media, and influencer marketing;
- Industries in which digital advertising is a source of revenue - search, social media and other online content, online newspapers, online radio and on-demand TV; and
- Jobs supported by the businesses paying for advertising, and benefiting from the return on ad spend (ROAS) in terms of greater sales.
Digital advertising as a source of revenue
Using ONS data on the ratio of GVA to turnover in different industries, and Annual Business Survey data on turnover per job in each industry, Public First quantified what this revenue uplift meant in terms of jobs and GVA.
Here, Public First adopted broadly the same approach as the Advertising Association\u2019s original Advertising Pays report. Depending on industry, this involved adopting one of two approaches, or in some cases a mix of these:
- A revenue-attribution approach - drawing on individual company financial data on digital ad revenues as a share of total revenue to establish the share of jobs supported. E.g. if 80% of company revenues are from digital advertising, 80% of the jobs are said to be supported by digital advertising.
- A value-added approach - where such a breakdown of individual company finances is not available, data from the ONS Annual Business Survey and ONS Business Register and Employment Survey were used to establish the amount of jobs supported per million pounds of revenue in each industry. Multiplying this by digital ad spend data gives a number for jobs supported.
AA/WARC data was used to establish digital ad revenues by sector. This approach was used for online social media platforms, audio, video on demand, newspapers, Connected TV, retail media and influencers.
Jobs supported in digital advertising
Digital advertising also supports jobs for the businesses paying for advertising space - through the additional sales that advertising brings in - i.e. the return on advertising spend (ROAS). There are numerous estimates of ROAS. Public First drew on research by Thinkbox which suggests an average ROAS of 3.33. This was higher than the ROAS used in the 2023 Digital Dividend report. Using AA/WARC data on digital ad spend, and IAB UK data on the industries paying for advertising, Public First estimated the increased revenues for businesses from the ROAS.
Indirect GVA and jobs
Public First\u2019s headline numbers also include GVA and jobs supported along supply chains (\u201cindirect impacts\u201d). To do this, Public First used its in-house economic model which draws on ONS \u201cinput-output\u201d tables which detail the interrelationships between different parts of the economy.
GVA and jobs forecasts
The GVA direct value estimates for the sectors: ROI in advertising, ad tech, connected TV, retail media and influencer marketing are used as the baseline for the forecast. Where existing IAB research has been conducted on growth in these sub-sectors such as in the IAB 2025 Adspend report and growth of retail media, these estimates have been used in constructing the baseline. The growth rates for ad tech, connected TV and influencers have used external forecasts.
The growth rates for each subsector are assumed to remain the same between 2025 and 2035. This is a reasonable assumption given they are a fast growing sector and that downside or upside assumptions cannot be justified without more information into the growth trajectory of the UK or sector.
The two policy environments are developed as follows:
- A supportive policy environment is estimated by looking at the fastest growing countries in the sector, like the US. This growth rate, which is higher than the combined baseline growth rate, is used as the supportive policy scenario.
- An unsupportive policy environment is estimated based on the growth rate of UK peer countries with lower growth but a similar economic landscape. This is similar to EU countries such as Germany and France. The digital advertising growth rate in these countries is still high but lower than the baseline.
The same growth rate is assumed between the GVA forecast and the growth rate for sectoral jobs. This is due to limited data on the difference between growth in GVA and the growth in jobs within these sectors.
Creator growth
Public First used global and UK influencer advertising spend data from Statista, available until 2030. Forecasts to 2035 are extrapolated using existing data, and establishes influencer ad growth over the next decade. ROAS estimates of 3.33 produced by Thinkbox were used to estimate the return on additional ad sales.
Chapter 2
Economic value of innovations outside London
Public First used estimates from their economic impact assessment on the value of ROI from digital advertising as the baseline. Regional estimates on the size of the advertising sector, produced by BiGGAR Economics, is used as the basis to extract the contribution of London and regions outside of London to digital advertising.
IAB survey data stating the share of companies currently using tools such as agentic buying is used as an approximation for the distribution of agentic buying across regions. Public First then multiplied this figure by the average ROAS figure of 3.33 produced by Thinkbox. The difference between the baseline value and the impact post-ROAS is the value added due to innovations such as agentic buying.
City job clusters
Public First defined digital advertising clusters as London, Manchester and Leeds, where the sector contributes to almost 50% of activity across the UK. The cities are taken from analysis in Advertising Pays, which estimates the concentration of jobs through LinkedIn Talent data. GVA values for the fastest growing areas such as ad tech, connected TV and retail media are taken from the economic impact model, as current values and estimates to 2035. These differences are apportioned across the three city clusters and the difference subtracted, over the next decade.
Regional GVA impact
Public First uses the estimates from their economic impact assessment on the GVA and jobs impact of digital advertising over the next decade. The cities are based on analysis from Advertising Pays, which estimates the concentration of jobs through LinkedIn Talent data. The GVA and jobs growth is apportioned using this concentration ratios across each city.
Ad tech accelerator value
Public First used Data City estimates for the total number of ad tech companies in the UK, adjusted with data from Beauhurst estimating the share of ad tech firms based outside London. Company growth rates - those growing above 20% a year - are taken from Data City, and defined as a scale up in this analysis. The value of an accelerator programme is the median per business value of two programmes: Digital Growth Grant and UKRI\u2019s Innovator Accelerator. The additional investment generated from the ad tech accelerator programme is based on evaluation of the Creative Industries Cluster programme.
Chapter 3
AI-augmented personalised advertising
Public First used the IAB 2025 Adspend report to determine the sector\u2019s current level of ad spend. Evidence across the various channels used for digital advertising suggests on average return on ad spend (ROAS) multiplies by 3 for each pound spent. AI effectiveness is estimated using research by McKinsey. Multiplied together the uplift on ad spend due to the increased effectiveness of AI is estimated to be \u00a312bn for the sector.
AI uplift on digital advertising
The estimate on the effectiveness of AI in increased return on ad spend is used as a foundation. Industry estimates suggest that AI improves advertising effectiveness through better targeting, automated optimisation, and enhanced bidding and creative generation. Ad effectiveness is used to estimate the full potential impact of AI, which looks at two avenues: 1) ROAS uplift value and 2) the AI ROI effect.
ROAS uplift value used an average ROAS benchmark from the literature (~\u00a33 in this case). The AI uplift is based on research by McKinsey, using the lower bound of 10%, on the return on investment generated by adopting AI tools. Since this figure estimates at its full potential AI-advertising could generate a 10% uplift in ROAS, the scenario assumes this 10% can be reached within the 10 year period when AI adoption is widespread.
Value of agentic buying
Public First used estimates from their economic impact assessment on the value of ROI from digital advertising. ROAS estimates of 3.33 produced by Thinkbox were used to estimate the return on additional ad sales, and multiplied by the impact of agentic buying on ROI. IAB survey data is used to determine the share of businesses who are not currently using agentic buying, and will benefit from the productivity and sales uplift.
Chapter 4
Negative shock to advertising sector
Public First used the IAB\u2019s 2025 Adspend figures to determine the sector\u2019s current level of ad spend. A negative policy shock is assumed to impact 34% of digital advertising and research from Goldberg et al. (2024) estimates the reduction in website views of 12% is used.
PECR reforms
Public First assumes that changes to PECR would increase cookie collection by 22%, leading to an increase in ad spend. The analysis is restricted to the use of UK adults, ages 16 and above, of whom 18% reject cookies on a daily basis. The analysis is primarily limited to open display digital adverts, as this medium uses cookie-based targeting most, and makes up 14% of all digital ad spend.