What Do Corporate Wellness and Fitness Programmes Actually Return?

Corporate wellness ROI is the most over-claimed number in human resources. Vendors quote returns of three to six dollars per dollar; the best randomised evidence says close to zero on healthcare costs; and CEOs in Wellhub’s 2025 survey still report positive returns by an 82% majority. All three can be true at once, because they measure different things over different horizons. This article separates them, using the 2024–2026 evidence on what fitness, sport and broader wellbeing programmes actually return, and how a finance team should model it. It is part of the Sports Economy hub and complements our review of whether workplace sport improves productivity and sports sponsorship as an employer-branding tool.

How big is the corporate wellness market, and what are employers buying?

Grand View Research values the global corporate wellness market at $55.1bn in 2025, forecasting $56.7bn in 2026 and $70.1bn by 2033 at a 3.1% compound rate. North America accounts for 39%, large organisations for 53.3% of spend, and health risk assessments remain the single biggest service line at 21%.

The growth rate is the telling number. A market compounding at 3% is a mature one, and the mix is shifting from screening-led programmes towards stress management, fitness aggregators and mental-health platforms, the segments Grand View expects to grow fastest. The sport and fitness slice sits inside this: subsidised gym and class access through aggregators such as Wellhub (formerly Gympass, which passed 250 million check-ins on its network in 2025), on-site fitness, corporate sports leagues and step-challenge platforms. Wellhub’s own 2025 Work-Life Wellness survey of more than 5,000 employees reports that 85% rate wellbeing as important as salary and 79% would consider leaving an employer that does not prioritise it, which is the demand signal vendors sell against.

What employers pay varies by format: aggregator subscriptions cost $20–60 per active user per month, on-site facilities several hundred dollars per head a year, a corporate league entry tens of dollars per participant. The cost side is easy to establish. The problem is the benefit side, where three decades of research have produced three incompatible answers.

What did the randomised trials find?

Almost nothing on cost. The two largest randomised evaluations, Song and Baicker’s BJ’s Wholesale Club trial published in JAMA in April 2019 and the Illinois Workplace Wellness Study by Jones, Molitor and Reif, found no significant effect on medical spending, absenteeism, tenure or job performance within 18 to 24 months.

The BJ’s trial is the gold standard because of its design: 160 worksites in the eastern United States, 20 randomly assigned to receive an eight-module programme on nutrition, physical activity and stress run by registered dietitians, 140 as controls; 4,037 treated employees and 28,936 controls. After 18 months the treated sites showed two real behavioural changes, an 8.3-point rise in regular exercise and a 13.6-point rise in active weight management, and no significant effect on 27 other self-reported health measures, 10 clinical markers including BMI, blood pressure and cholesterol, 38 spending and utilisation measures, or on absence, tenure and performance. The authors’ three-year follow-up, published in Health Affairs in 2021, reported the same absence of clinical or spending effects.

The Illinois study reached the same conclusion by a different route. Of 12,459 benefits-eligible university employees invited, 4,834 completed the survey and 3,300 were randomised into six treatment arms of the iThrive programme (biometric screening, health risk assessment and wellness activities, with incentives of $50–350 in year one). There was no effect on medical spending, absenteeism or productivity after two years. Two findings matter more for sport and fitness programmes specifically. First, healthier employees enrolled at higher rates: the people the programme reached were those who needed it least. Second, incentives had steeply diminishing returns: $100 lifted screening completion from 47% to 59%, and doubling it to $200 added only four more points.

Harvard Medical School’s summary of the BJ’s result notes that 18 months may be too short for health benefits to show. Even so, the implication for a finance director is clear: the healthcare-savings line in a vendor’s ROI model should be zero for at least two years, and any benefit in that window has to come from somewhere else.

Why did earlier studies report returns of 3:1 or more?

Because they were not randomised. Baicker, Cutler and Song’s 2010 Health Affairs meta-analysis of 36 mostly observational studies found $3.27 saved in medical costs and $2.73 in absence costs per dollar spent, about 1.8 days per employee a year. Those figures came from participants compared with non-participants, which bakes in selection bias.

The same authors who produced the 3:1 estimate went on to run the BJ’s trial, and their later work is the most credible correction in the field: once you randomise who gets the programme, the healthcare savings disappear. Observational studies compare people who chose to join with people who did not, and the joiners are fitter, more engaged and less likely to be sick in the first place. The Illinois study documented this directly with its positive-selection finding, and warned that programmes tied to premium discounts can shift costs onto lower-income and sicker employees who do not participate.

RAND’s seven-year evaluation of PepsiCo’s Healthy Living programme, covering more than 67,000 eligible workers and published in January 2014, is the bridge between the two literatures. It split the programme into its components and found that disease management (structured support for employees with diagnosed chronic conditions) cut costs by $136 per member per month, a return of $3.78 per $1, while lifestyle management (fitness, nutrition, smoking cessation) did not generate savings above its own cost. Employees in both components saved $160 a month, almost all of it from the disease side. The lesson is not that wellness fails; it is that returns concentrate in the small, sick, expensive population and not in the broad, healthy one that gym subsidies reach.

What does the 2024–2026 evidence say about fitness and mental-health programmes?

It splits sharply. Oxford’s William Fleming found in 2024 that individual-level interventions such as mindfulness apps and resilience courses produced no wellbeing gain across 46,336 UK workers, while Deloitte’s 2024 report puts organisation-wide mental-health measures at £6.30 per £1. Structural programmes beat apps.

Fleming’s study, published in the Industrial Relations Journal in January 2024, used Britain’s Healthiest Workplace data across 233 organisations to compare employees who used 11 popular wellbeing offers with colleagues who did not. None of the individual-level practices (mindfulness, resilience and stress-management training, relaxation classes, wellbeing apps, coaching) showed a benefit, and several showed a small negative association, which Fleming attributes to people with lower wellbeing being more likely to take them up. His conclusion is that employers need “more ambition”: changes to workload, scheduling, autonomy and management practice rather than another app licence.

Deloitte’s figures point the same way from the cost side. Its 2024 UK analysis estimates poor mental health costs employers about £51bn a year, of which presenteeism is £23.8bn, turnover £19.5bn and absence £7.3bn, and that the average return on mental-health investment is £4.70 per £1, rising to £6.30 for universal, organisation-wide interventions and falling to £4.10–4.20 for proactive and reactive individual support. Screening returns £5.90, training and awareness £5.00, therapy £3.30. These are modelled rather than trial-based returns, so treat them as a planning range, but their ordering (culture and structure first, individual treatment second) matches Fleming’s trial-adjacent evidence.

Where does sport sit? It is the one individual-level activity with a plausible structural component, because team formats change the social fabric of a workplace in a way an app cannot, and because physical activity acts on both the musculoskeletal and the mood channels of presenteeism at once. The randomised trials did not test team sport specifically, so the fair position is that sport’s return is likely to resemble Deloitte’s organisation-wide estimate when it is run as an in-hours, team-based, widely attended programme, and the PepsiCo lifestyle estimate (below cost) when it is run as a subsidised gym membership taken up by the already fit.

Why do CEOs still report a positive return?

Because they count retention and engagement, which the trials did not. Wellhub’s 2025 Return on Wellbeing study of 1,500 CEOs and leaders found 82% see positive ROI, 78% see returns above 50% and 30% above 100%; the top driver cited was stronger retention (73%), ahead of healthcare costs (68%).

The gap between the executive surveys and the trials is a measurement gap, not a contradiction. The JAMA and Illinois studies measured claims data, clinical markers and HR records over 18–24 months; they were not designed to capture whether a programme made a 28-year-old engineer slightly less likely to accept a recruiter’s call. Yet turnover is the second-largest line in Deloitte’s cost model at £19.5bn, and in professional services replacing an employee typically costs six to nine months of salary. If a wellbeing programme that costs $300 per head a year reduces regretted attrition by even one percentage point in a 1,000-person firm, the ten avoided departures are worth more than the $300,000 programme on their own.

Wellhub’s earlier 2023 survey of more than 2,000 HR leaders across nine countries found 90% of companies that measure their programmes report positive ROI. Both surveys are vendor-commissioned and self-assessed, so they are not evidence of return; what they show is where executives believe the return sits, and that order (retention first, then productivity, then healthcare) is the right one to build a business case around.

What should a sport or fitness programme actually return?

Modelled conservatively, a well-targeted sport or fitness programme returns roughly 1.2–2.5 times its cost through presenteeism and retention within two years, with healthcare savings as an unbudgeted upside. Poorly targeted programmes that subsidise the already active return less than their cost, exactly as RAND found for PepsiCo’s lifestyle tier.

Worked example. A 1,000-person professional firm with a loaded cost of $90,000 per employee runs an in-hours team-sport and movement-break programme at $120 per head, or $120,000 a year. Vitality’s Britain’s Healthiest Workplace data suggests about 37% of staff are inactive; if the programme converts a quarter of them, around 90 employees, and each recovers three productive days a year (the lower end of RAND Europe’s up-to-five-days estimate for moving from inactive to WHO-recommended activity), that is 270 days, or roughly $100,000 of output at a $370 daily cost. Add one percentage point less regretted attrition, ten people at $45,000 replacement cost each, and the retention layer is worth $450,000. Total modelled benefit: about $550,000 against $120,000, before any healthcare effect. Halve every assumption and it still pays.

Now run the same arithmetic for a $600-per-head on-site gym used by 150 staff, of whom 120 already trained elsewhere. The programme reaches 30 previously inactive people, recovers perhaps 90 days ($33,000), and its retention effect is concentrated in a cohort that was already engaged. Cost $600,000, benefit well under $200,000. Same company, same evidence base, opposite result. The ROI of corporate fitness is a design variable, not a property of fitness.

Türkiye offers a cheap illustration of the first model: a company team in the Istanbul Corporate Games, which holds its 24th edition across 15 sports in June 2026, costs entry fees and a few hours of paid time, and reaches employees who would never use a gym. Japan’s METI certifies the output: 3,400 large companies and 19,796 SMEs held Health and Productivity Management recognition in 2025, a disclosure Japanese institutional investors read as a governance signal.

How should a finance team audit a vendor’s ROI claim?

Ask four questions: was the comparison randomised or at least matched on baseline health; what was the participation rate among previously inactive staff; over what horizon were healthcare savings measured; and how much of the claimed return is retention that the vendor cannot attribute to itself. Most claims fail the first two.

The vendor ROI calculator is a sales tool, and its usual method, comparing participants with non-participants, is exactly what the Baicker and Illinois work shows overstates return. Request the raw participation data instead: the share of enrolled users active each month, broken down by baseline activity, and 12-month user retention. A platform that cannot tell you whether its users were already gym members cannot tell you its return.

Benchmark against the public evidence rather than the vendor’s cases: Deloitte’s £4.70 is a reasonable ceiling for any wellbeing line, RAND’s below-cost lifestyle finding a reasonable floor for untargeted fitness subsidies, and the trials set healthcare savings to zero for two years. Then insist on a measurement plan before signing. A phased launch across offices or cohorts gives you a control group for free.

Frequently Asked Questions