Services
Six core services, and how to tell which one you need
Each one is described the same way: the decision, how Ensemble approaches it, and what you get, along with common questions and limits.
We scope each engagement to the smallest effort that can credibly answer the question. A founding partner does the work directly, focusing their time where expertise matters most: framing the problem, choosing the right approach, interpreting the evidence, and assessing uncertainty. They stay available to discuss the findings and what they mean for your decision.
Start here
Start with the decision, not a generic risk score
The questions clients arrive with. If one is close to yours, follow it to the service that answers it.
- Which facilities or suppliers face the greatest changes in heat, drought, flood, or water availability?Climate and water risk →
- Can weather risk be hedged with a derivative or parametric product, and how should the index be designed?Weather-risk products and hedging →
- Which candidate locations face the least climate and water risk over the life of the asset?Site selection analysis →
- How often has an event like this season's occurred here, and is it becoming more frequent?Forecasting and historical analytics →
- Is an existing climate model or vendor analysis scientifically fit for its intended use?Independent technical review →
- How should uncertainty be communicated to executives, investors, customers, or public stakeholders?Climate and water risk →
- Would this proposal survive a hostile technical review?Roadmaps and proposals →
01
Site selection analysis
The decision
Choosing where to build, grow, buy, or invest can shape exposure for decades. We support decisions ranging from facility and infrastructure siting to agricultural sourcing, resource availability, and investment screening.
How Ensemble approaches it
We bring together the environmental, geographic, and socioeconomic information relevant to the decision, from climate, hazards, water, land, and natural resources to infrastructure and workforce, and compare candidate locations against the criteria that matter.
What you get
A clear, defensible comparison of candidate locations: where they differ, what drives those differences, and how confident you can be in the result.
Common questions
- Which candidate locations carry the least risk across climate, water, energy, and workforce over the life of the asset?
- Where are water, land, and energy resources sufficient for what we plan to build or grow?
- How do candidate sourcing regions compare on yield stability and weather risk?
- Which site-level findings are material enough to change the choice?
Important limitations
- A site assessment informs the comparison; it is not an engineering, geotechnical, legal, or permitting evaluation.
- Where candidates cannot be credibly distinguished on the evidence, we will say so rather than manufacture a ranking.
Illustrative deliverables
Illustrative Ensemble analysis
Which sourcing regions stay reliable as the climate shifts? Share of years each hazard hit damaging levels at each crop stage, 1991–2020. Synthetic data. Illustrative Ensemble analysis
ensemble_results_v1.csv: excerpt
Typed results with units, scenario, horizon, and ensemble range
Sample rows from a delivered results file for fictional sites. site_id metric scenario horizon value units p10_p90 SITE-0142 days_tmax_gt_35c SSP2-4.5 2050 41.6 days yr⁻¹ 28.4 – 57.1 SITE-0142 annual_runoff SSP2-4.5 2050 512 mm yr⁻¹ 398 – 641 SITE-0207 rx1day_100yr SSP5-8.5 2050 186 mm 151 – 233 SITE-0207 spei12_below_-1.5 SSP5-8.5 2050 0.18 fraction of months 0.09 – 0.31 Accompanied by a data dictionary, a method reference for every metric, and a checksum manifest.
How do results plug into your systems? Every number arrives typed, with units, scenario, horizon, and an uncertainty range. Synthetic data.
02
Climate and water risk
The decision
A risk committee, regulator, investor, or planning process is asking how climate and water conditions could affect your assets and operations. The answer has to hold up to scrutiny, and it has to say which uncertainties actually matter to the decision.
How Ensemble approaches it
We start from the decision or report the analysis supports, then build the evidence beneath it: historical baselines, projections across scenarios, and uncertainty analysis, interpreted by the scientists who produced them.
What you get
Defensible analysis of the risks that matter, documented methods and limitations included, in the form your audience needs: a report, maps and tables, data, or a briefing.
Appropriate precision, and no more
We use only the specificity the science supports. A convincing number at the wrong resolution invites decisions the analysis cannot carry. Where the data will not support the detail, we say so.
Specificity is justified by
- The source data
- The model
- The uncertainty
- The intended decision
- The analytical method
Common questions
- Which facilities or suppliers face the greatest changes in heat, drought, flood, or water availability?
- How should physical climate risk be represented in an internal or external report?
- Where would better climate information change a decision rather than simply confirm existing knowledge?
- How should uncertainty be communicated to executives, investors, customers, or public stakeholders?
Important limitations
- Projections describe plausible ranges under stated scenarios and assumptions, not predictions of what will happen at a specific place and time.
- A credible outcome may be that better climate information would not change the decision. We will say so.
Illustrative deliverables
Illustrative Ensemble analysis
Change in hazard-exposure days per year by 2050 vs a 1991–2020 baseline, mid-range scenario. Hatching: under 70% ensemble agreement on sign. Synthetic portfolio. Illustrative Ensemble analysis
Annual runoff, mm yr⁻¹: observed record to 2024, then the 10th–90th percentile of a 20-member ensemble. Dashed line: operating threshold. Synthetic data. Illustrative Ensemble analysis
Line weight shows the hypothesized strength of each hazard-to-decision link. Links are hypotheses to test, not measured relationships. Fictional example.
03
Forecasting and historical analytics
The decision
An operational or planning decision depends on conditions ahead, or on how often something has happened before: a frost at flowering, a low-inflow season, a demand spike. Off-the-shelf products rarely fit the question.
How Ensemble approaches it
We build what the question requires: historical and event-frequency analysis, forecasts tuned to your lead time and locations, and the validation that shows whether the skill is real. Everything we build is documented and handed over.
What you get
A fit-for-purpose analysis or forecasting workflow your team can run, with the verification to know when to trust it, and an honest answer when skill at your lead time is not there.
Built for your workflow, then handed over
One size fits nothing in the sectors we serve; even within a single industry, no two workflows look alike. We build the solution to fit yours, train your team to operate and maintain it, and hand it over, faster and leaner than an institutional consultancy, with no junior bench and no overhead to fund. We have built an AI-native platform before, so we know what that takes, and when you don't need one.
Common questions
- How often has an event like this season's occurred here, and is it becoming more frequent?
- What forecast or projection horizon is actually useful for the decision?
- Can reservoir inflows, crop yields, or energy demand be forecast with usable skill at our lead time?
- Which past years are the best analogs for the conditions ahead?
Important limitations
- Where skill is low at the lead time you need, we will report that rather than present a model that looks confident.
- Models we build are documented and handed over; we do not require you to keep buying a platform to use them.
Illustrative deliverables
Illustrative Ensemble analysis
Storage as % of usable capacity over 12 months: median with 25th–75th and 10th–90th percentile bands of a 40-member inflow ensemble. Synthetic data. Illustrative Ensemble analysis
24-hour rainfall depth, mm, by return period (log axis): observed maxima, fit with 90% CI, and a mid-century baseline. Synthetic data.
04
Weather-risk products and hedging
The decision
Some weather risk cannot be engineered away and has to be carried or hedged. Whether you are building a weather-linked product or deciding whether to buy one, the question is whether an index can credibly track the exposure.
How Ensemble approaches it
We combine the atmospheric science that defines a sound index with market experience in structuring and pricing: exposure and driver analysis, index and trigger design, and hedge-effectiveness evaluation against history and simulation.
What you get
An index and trigger specification with its pricing analysis and basis risk stated plainly, documented for counterparties, and a straight answer if the exposure cannot be hedged well.
Common questions
- Can this weather exposure be hedged with a derivative or parametric structure?
- How should the index be defined so payouts track actual losses?
- What basis risk remains, and is it acceptable for the decision?
- Is an existing weather-linked product priced and structured defensibly?
Important limitations
- We design and evaluate the science and structure of a product; we are not a broker, insurer, or investment advisor, and we do not execute trades.
- Where an index cannot track the exposure tightly enough to hedge it, we will say so rather than engineer false comfort.
05
Independent technical review
The decision
You are being asked to rely on climate analysis you did not produce: a vendor's model, a consultant's report, a forecast feeding a decision. Someone qualified and independent needs to say whether it is fit for the purpose.
How Ensemble approaches it
We evaluate the methods, data, baselines, scenarios, and claims against the intended use, in writing, with the same scrutiny we would apply to our own work.
What you get
A written technical review you can put in front of a board, a regulator, or the vendor: what is robust, what is uncertain, and what the analysis can credibly carry.
Common questions
- Is an existing climate model or vendor analysis scientifically fit for its intended use?
- Are the methods, baselines, and scenarios defensible for the reporting context?
- What is the simplest analysis that can answer the question credibly?
- What in this result is robust, what is uncertain, and what can it actually carry?
Important limitations
- A review assesses whether a method is fit for a stated purpose, not a certification, an assurance opinion, or a warranty of anyone's results.
- Findings are written to be shared with the party under review; we do not write conclusions we would not defend in front of them.
Illustrative deliverables
Illustrative Ensemble analysis
Technical review memorandum: excerpt
Fitness-for-purpose assessment: Northwind Analytics physical-risk report
Claims assessed, findings, and whether each claim is supported. Claim assessed Finding Assessment 100 m effective resolution for heat metrics Downscaling inherits an effective resolution nearer 8 km. Subgrid detail is interpolated, not resolved. Not supported Baseline period 1991–2020 used throughout Baseline is consistent across hazards and matches the reporting period. Documented clearly. Supported Flood hazard reflects local drainage Drainage is not represented. Suitable for screening; not for site-level design decisions. Partly supported Ensemble spread represents model uncertainty Spread covers model uncertainty but not scenario or internal variability. Ranges are understated. Partly supported Each row: the claim assessed, the finding, and whether the method supports the stated use. Fictional example. Illustrative Ensemble analysis
Observed vs modeled with the 1:1 line; skill by lead time vs climatology, including the lead where skill is not usable. Synthetic data.
06
Roadmaps and proposals
The decision
A program needs a technical plan that holds up, and often the funding to build it. The roadmap has to be credible in delivery, and the proposal has to win.
How Ensemble approaches it
We define or refine the roadmap with scope, sequencing, and risks, and build the proposal's technical case, drawing on experience writing, winning, reviewing, and delivering these documents on both sides of the table.
What you get
A roadmap or proposal that survives hostile technical review, and a plain statement beforehand if the plan the evidence supports is smaller than the one you hoped for.
Common questions
- Is this roadmap technically credible, and in the right order?
- What would make this proposal fundable?
- Can our draft survive a hostile technical review?
- What program should we propose, given what we can actually deliver?
Important limitations
- We draft and strengthen the technical case; we are not a proposal mill, and we take work only in domains where we can stand behind the substance.
- If the roadmap the evidence supports is smaller than the one you hoped for, we will say so before you submit.
Next step
Not sure which of these fits?
Describe the decision or reporting need in your own words. Part of our job is telling you which engagement, if any, is the right size for the question.
Describe your project