How the Offsite ROI Calculator works

What we price, what we refuse to price, and a fully worked example. For the person who has to defend the number.

Try the calculator →

The Offsite ROI Calculator estimates what stronger team connection is worth per year for a remote or distributed team. This page explains how, in the same words we use with any advisor or CFO who asks. Every figure here restates what the emailed report already discloses, with its source.

The rules the calculator lives by

  1. Only four things ever get a money value: the people you keep, extra productive capacity, recovered connection days, and AI upskilling. Each is priced on its own distinct mechanism, so nothing is counted twice. Everything else, however impressive, is shown as supporting evidence with its citation and is never quietly added to your total.
  2. We never claim the whole gap. The engagement research compares top-quartile and bottom-quartile teams and is correlational. You choose a scenario: conservative, realistic or ambitious (capture 25%, 50% or 75% of the documented gap), and every engagement-based line scales by your choice.
  3. Verified at the primary source, or not used. Every statistic was checked on the publishing organisation's own page or PDF. Anything we could not verify went on a rejected list; a sample is below.
  4. Ranges come from the sources' own published ranges, never from invented spreads, and the low end is the only number we ever lead with.
  5. We say what kind of euros they are: avoided costs, capacity value and recovered time, not guaranteed cash flow. Payback is value payback, not cash-flow payback.
  6. The math rounds against us. Payback months round up, the breakeven line prices a departure at the lowest documented replacement cost, and an overpriced program honestly shows a negative return.
  7. Priced for the people in the room. In a large organisation, read the numbers per attending group, not for the whole company.

The four priced lines

1. People you keep (avoided cost). Gallup's Q12 meta-analysis (183,806 business units, 3.35 million employees) finds 21% lower turnover in top-quartile engagement teams; Gallup's published gap for low-turnover organisations is actually 51%, so using 21% understates rather than overstates. Each prevented departure is valued at 0.9 to 1.0 times salary, while the SHRM literature documents total turnover costs of 0.9 to 2.0 times salary. We price only your chosen share of that 21% gap.

2. Productive capacity unlocked (capacity value). The same Gallup meta-analysis finds top-quartile teams 14% more productive on production records (18% on sales; we use the lower figure). Valued at salary cost, never at revenue, and scaled by your scenario.

3. Connection days recovered (recovered time). Gallup's State of the Global Workplace 2026 (263,810 respondents) measures that 24% of remote and 24% of hybrid employees feel lonely every day, and Cigna's loneliness research finds lonely workers miss roughly 5 more workdays a year. We price only the recovered days for the lonely fraction of your remote and hybrid people, at daily salary cost, scaled by your scenario. Office-first teams get a strongly reduced share.

4. AI-native upskilling catalyst (capacity value, conditional). A peer-reviewed field study in The Quarterly Journal of Economics (5,172 support agents) measures that access to a generative-AI assistant lifts productivity by 15% on average and 30% for less-experienced workers. This line only appears when your team is not yet AI-native, and we attribute only a deliberately conservative share of the measured annual gain to the offsite, because an upskilling sprint is the catalyst of adoption, not the whole journey.

A worked example

These are the numbers the live calculator produces for this profile; you can reproduce them at driftawave.com/roi-calculator.

Team size40 people
Average fully-loaded salaryEUR 70,000
Voluntary attrition13% (the EU default, editable)
Work modelHybrid and remote-first
AI adoptionExperimenting, not yet embedded
Program1 offsite per year, 3 days, EUR 1,500 per person all-in
ScenarioRealistic (capture 50% of the documented gap)
Cash investmentEUR 60,000
True cost including the team's 3 gathering daysEUR 98,182
People you keepEUR 34,398 to 38,220
Productive capacity unlockedEUR 196,000
Connection days recoveredEUR 7,636
AI-native upskilling catalystEUR 52,500 to 105,000
Total upsideEUR 290,534 to 346,856
Payback on cash investment~3 months
Payback on true cost~5 months

How each line is built with these inputs: 40 people at 13% attrition means 5.2 expected departures a year; the Gallup gap is 21% and the realistic scenario claims half of it, so 0.55 prevented departures valued at 0.9 to 1.0 times salary gives EUR 34,398 to 38,220. The salary base of EUR 2.8 million times the 14% Gallup productivity gap, at half, gives EUR 196,000. The lonely quarter of the team reclaiming half of 5 missed days at daily salary cost gives EUR 7,636. The AI line prices the measured 15 to 30% gain, of which we attribute only a deliberately conservative share to the offsite as catalyst, further reduced because this team is already experimenting: EUR 52,500 to 105,000. Payback divides the investment by the low end of the upside and rounds up, against us. The exact attribution and calibration factors are part of our internal model; on a call we run this same walkthrough on your own numbers, line by line, source by source.

Context that is never added to the total: at 13% attrition this team already loses 5.2 people a year, which costs EUR 327,600 to 728,000 in replacement costs at the documented range. The whole program costs the equivalent of replacing one person.

What we refused to count

  • "5% better retention means 25 to 95% more profit": that finding is about customer retention (Reichheld and Sasser, 1990); applying it to employees is a category error.
  • Gallup's own 23% profitability gap: real and verified, but shown as evidence only, because pricing it would double count with the productivity line.
  • A priced "faster decisions" line: no trusted source measures how much an offsite improves decision speed, so it stays evidence.
  • An "AI-native premium": organisation-level comparisons of AI winners versus laggards are correlational with selection effects, so they carry the story, never the price.

Every source, with its caveat

The complete evidence base: 26 verified sources. Six of them power the priced lines; the rest are shown in your report as supporting evidence and are never added to your total. Each carries the honest caveat we would want to hear if we were the buyer. Defaults are calibrated against 80,000+ public labour-market data points (BLS, Eurostat, OECD, CBS, Statistics Canada, ABS).

Impact: retention and the cost of leaving

  1. Total costs of turnover range from 90% to 200% of the annual salary (direct replacement costs alone reach 50-60%).SHRM Foundation, "Retaining Talent: A Guide to Analyzing and Managing Employee Turnover" (D.G. Allen) (2008) · Tier A · used in the mathSHRM documents total turnover costs of 0.9x to 2x of annual salary. The calculator caps its own math at 0.9x to 1.0x of salary; the full range is shown for context only.
  2. Top-quartile engagement teams see 21% lower turnover than bottom-quartile (51% lower in low-turnover organizations).Gallup Q12 Meta-Analysis, 11th edition (183,806 business units, 3.35M employees) (2024) · Tier A · used in the mathCorrelational quartile comparison, not a guaranteed effect of any single intervention. The calculator only claims the share of this gap you choose as your ambition level, and uses the 21% figure, not the 51% one.
  3. Top-quartile engagement business units are 23% more profitable than bottom-quartile.Gallup Q12 Meta-Analysis, 11th edition (2024) · Tier A · evidence onlyCorrelational; shown as evidence, deliberately kept out of the totals to avoid double counting with productivity.
  4. Top-quartile engagement business units score 10% higher on customer loyalty and engagement.Gallup Q12 Meta-Analysis, 11th edition (2024) · Tier A · evidence onlyCorrelational; evidence display only. Customer-NPS gains are real but not additively priced into the totals.
  5. Low engagement costs the global economy $10 trillion, or 9% of global GDP; only 20% of employees are engaged (2026, the lowest since 2020).Gallup, State of the Global Workplace 2026 (2026) · Tier A · evidence onlyMacro-economic extrapolation; context framing only.
  6. Burned-out employees are 2.6x more likely to be actively seeking a different job and 63% more likely to take a sick day.Gallup, "Employee Burnout, Part 1: The 5 Main Causes" (2018) · Tier A · evidence onlyRisk framing for the "cost of doing nothing" panel only.
  7. In a randomized controlled trial, one office day a month for fully-remote employees raised productivity by 7.8%, added 36 minutes of colleague communication in the following week, and cut attrition by a third, with a benefit-cost ratio of about 5 to 1.Aksoy, Bloom, Davis, Marino & Ozguzel, "The Value of One Office Day a Month", NBER Working Paper 35331 (2026) · Tier B · evidence onlyWorking paper (not yet peer reviewed); one firm, customer-service work, and one office day is not a multi-day retreat. The direction transfers, the numbers do not: we price nothing on this and we do not sell office days.

Execution: productivity and AI

  1. Top-quartile engagement teams are 14% more productive (production records; 18% on sales).Gallup Q12 Meta-Analysis, 11th edition (2024) · Tier A · used in the mathCorrelational quartile comparison. The calculator uses the lower 14% figure and multiplies by your chosen ambition level, valuing capacity at salary cost, not revenue.
  2. Access to a generative-AI assistant increased worker productivity, measured as issues resolved per hour, by 15% on average, and by 30% for less skilled and less experienced workers.Brynjolfsson, Li & Raymond, "Generative AI at Work", The Quarterly Journal of Economics 140(2) (peer-reviewed; 5,172 support agents) (2025) · Tier A · used in the mathPeer-reviewed field study in one support organization; fieldwork ran November 2020 to February 2021 on a GPT-3-generation assistant, so published 2025 does not mean measured 2025. The low end of the priced range uses the 15% average, the high end the published 30% for less-experienced workers. The calculator applies it only when your team is not yet AI-native, discounted by an attribution factor because an offsite upskilling sprint is a starting catalyst, not the whole adoption journey.
  3. The most AI-exposed companies show 34% versus 24% productivity growth, and the wage premium for AI skills reached 62%.PwC, Global AI Jobs Barometer 2026 (one billion+ job ads, 27 countries) (2026) · Tier B · evidence onlyCorrelational, organisation-level exposure comparison (consultancy research, disclosed). Evidence for what is at stake at the organisational level; never priced.
  4. Cautious companies, 15% of the sample, are growing two percentage points more slowly than their peers, with profit margins three percentage points lower.PwC, 29th Annual Global CEO Survey (4,454 CEOs, 95 countries) (2026) · Tier B · evidence onlyCorrelational; the inaction framing: not moving is also a decision. Evidence only, never priced.
  5. AI high performers are nearly three times as likely to have fundamentally redesigned individual workflows; that intentional redesign has one of the strongest contributions to meaningful business impact of all factors tested.McKinsey, The State of AI in 2025 (1,993 participants, 105 countries) (2025) · Tier A · evidence onlySurvey of self-reported outcomes; high performers are ~6% of respondents. Evidence for why the offsite works on how the team operates, never priced.
  6. Organizational factors like culture, manager support, and talent practices account for more than 2x the reported AI impact of individual factors (67% vs 32%).Microsoft, Work Trend Index 2026 (20,000 knowledge workers) (2026) · Tier B · evidence onlyVendor research (Microsoft), disclosed as such. Evidence that AI value is built at the organisational level; never priced.
  7. Measured studies report generative AI output gains of 14 to 15% in customer support and 26% in software development, with smaller gains on work requiring deeper reasoning.Stanford HAI, AI Index Report 2026, ch. 4 Economy (synthesis of Brynjolfsson et al. 2025 and Cui et al. 2025) (2026) · Tier A · evidence onlyInstitutional synthesis, republished annually; the underlying studies measured 2020 to 2024 deployments. Less experienced workers tended to benefit the most.
  8. 39% of workers' core skills will change by 2030, and 59% of the workforce will need training before then.World Economic Forum, Future of Jobs Report 2025 (1,000+ global employers) (2025) · Tier B · evidence onlyEmployer expectations, not measured outcomes; context framing for the upskilling narrative only.
  9. In a pre-registered field experiment with 791 professionals, individuals working with a generative AI assistant matched the performance of two-person teams without one, and AI-assisted groups were 12 to 16% faster with a 9 percentage point higher chance of a top-decile solution.Dell'Acqua, Mollick et al., "The Cybernetic Teammate", Organization Science 37(4), field experiment at a global consumer-goods company (2026) · Tier A · evidence onlySingle firm, product-innovation tasks, data collected 2024. The same experiment finds AI also substitutes for part of the emotional benefits of teamwork - which is exactly why deliberate human connection needs its own investment. Cited whole, never halved.
  10. In nationally representative US data, workers using generative AI report time savings equal to 5.4% of their work hours, about 1.4% averaged across the whole workforce.Bick, Blandin & Deming, "The Rapid Adoption of Generative AI", Management Science (Federal Reserve Bank of St. Louis / Vanderbilt / Harvard, quarterly representative survey) (2025) · Tier A · evidence onlySelf-reported time savings, and self-reports are demonstrably unreliable in both directions; data from late 2024. Included deliberately as the population-level reality check next to the larger experimental gains.

Leverage: connection and collaboration

  1. Intentional in-person team gatherings boost team-connection scores by 27% on average, decaying back over ~4 months — pointing to roughly 3 gatherings a year.Atlassian Team Anywhere Lab (1,600+ team gatherings studied) (2024) · Tier B · evidence onlySingle-company internal research (Atlassian is fully distributed), large-N and longitudinal but not peer-reviewed. Used for cadence guidance and evidence, not multiplied into the totals.
  2. In-person pairs and teams generate 15-20% more creative ideas than videoconference ones; virtual teams are as good at selecting ideas, worse at generating them.Brucks & Levav, "Virtual communication curbs creative idea generation", Nature (lab + field study across 5 countries) (2022) · Tier A · evidence onlyMeasures idea generation in structured sessions, not offsites specifically. Peer-reviewed evidence display; pricing "ideas" in money would be speculative, so it stays out of the totals.
  3. 93% agree that a sense of belonging drives organizational performance.Deloitte, 2020 Global Human Capital Trends, "Belonging: from comfort to connection to contribution" (nearly 9,000 respondents, 119 countries) (2020) · Tier B · evidence onlySurvey agreement, not a measured effect size. Belonging overlaps the priced productivity and retention lines, so it is never priced; evidence display only.
  4. Firm-wide remote work made collaboration networks more static and siloed, with cross-group collaboration time dropping about 25%.Yang et al. (Microsoft), "The effects of remote work on collaboration among information workers", Nature Human Behaviour (61,182 employees) (2021) · Tier A · evidence onlyNatural experiment at one (very large) company during the 2020 remote shift; risk framing, not a priced line.
  5. 24% of fully-remote and 24% of hybrid employees experience daily loneliness (2026 edition).Gallup, State of the Global Workplace 2026 (263,810 respondents) (2026) · Tier A · used in the mathSelf-reported daily loneliness; Gallup 2026 measures the same rate for remote and hybrid workers, so the calculator applies it to both, and only to a reduced share for office-first teams.
  6. Lonely workers miss roughly 5 more workdays a year due to stress and sickness; workplace loneliness costs US employers an estimated $154B annually in stress-related absenteeism.Cigna Loneliness Index (n≈6,000; published in the Journal of Organizational Effectiveness) (2020) · Tier B · used in the mathUS pre-pandemic fieldwork; vendor research (Cigna is a health insurer), disclosed as such. The calculator prices only the recovered sick days for the lonely fraction of your remote and hybrid team, scaled by your ambition level.

Speed: hiring and decisions

  1. Filling a role takes a median 44 days and costs an average $4,683 per hire (executive hires average $28,329).SHRM Benchmarking, Talent Access Report (n=840 organizations) (2022) · Tier A · evidence onlyUS benchmark in USD, 2021 fieldwork. Shown as upside only: re-hiring costs are already inside the Gallup replacement-cost multiplier, so pricing them again would double count.
  2. 89% of talent professionals agree that measuring quality of hire will become increasingly important; 66% measure it through job performance ratings.LinkedIn, Future of Recruiting 2025 (surveyed talent professionals across 23 countries) (2025) · Tier B · evidence onlyVendor/platform research, disclosed as such; directional evidence for the A-game talent story, never priced.
  3. Decision effectiveness and financial results correlate at a 95% confidence level or higher in every country, industry and company size studied.Bain & Company, Decide & Deliver research (1,000+ companies over a decade) (2010) · Tier B · evidence onlyCorrelational, self-scored decision effectiveness, 2010. Evidence that speed and quality of decisions travel together; never priced.

If you want to challenge the numbers

Please do; the model is built for it. If your attrition is lower, if you believe the research does not transfer to your team, or if the program would not change how people work together, the lines shrink, and the calculator will honestly show that. Every line in your report carries a clickable citation to its primary source. Bring the report to a call and we will walk through any line against its source, with your numbers.

Calculate the ROI of your next offsite

Model last updated 27 July 2026; numbers recompute against the current model. Estimates are research-based projections, not guarantees.

Start Planning
Your Offsite

Let's create an unforgettable retreat experience together.

Get in Touch