Research that doesn't stop
at the data.
XandY is an independent research and strategy firm founded by two Yale scientists. We help organizations understand how people think and act, why they do, and what will actually move them.
It's best to do research backwards
Most research runs forward: collect the data, look for something interesting, then figure out what to do about it. That order feels natural, and it quietly wastes a lot of research budgets. The study ends, the findings are presented, and nobody can quite connect them to a decision.
We run the process in reverse, an approach borrowed from a classic Harvard Business Review argument for "backward" market research. We start at the end, with the actions you anticipate taking and what you would do differently depending on what the data shows. Sometimes that means sketching the final report before a single question is written. Working backward from the intended use tells us which insights you need, and the insights tell us which methods and data will get them. By the time a study fields, every question has a job.
A toolbox, not a menu
Most research providers lead with the method they sell. Polling shops run polls. Focus group facilities run focus groups. When a provider owns one tool, every question tends to get answered with it, whether or not it is the right tool for the job.
We work in the opposite direction. We start from your question and select the method that answers it best, because we have working fluency across the full range of quantitative and qualitative approaches. Just as important, we interpret results with each method's strengths and limits in plain view: what this data can support, what it cannot, and what to collect next if the decision demands more certainty.
How we collect
- National surveys & opinion polling census-matched samples, AAPOR-standard methods
- Tracking studies & longitudinal panels how attitudes move, not just where they stand
- Randomized experiments message testing, A/B designs, full RCTs
- Field experiments & program evaluation causal evidence from the real world
- In-depth interviews & focus groups the "why" behind the numbers
- Content & social media analysis what's being said, by whom, at scale
- Secondary data & evidence synthesis standing on what's already known
How we analyze
- Advanced statistical modeling multilevel models, causal inference, conjoint, SEM
- Audience segmentation & clustering who to talk to, in what order, with which message
- Text analytics & machine learning structure and signal from open-ended data
- Measurement design & psychometrics instruments that measure what they claim to
- Qualitative & content coding systematic, reliable, transparent
- Data visualization from static graphics to interactive explorers
With you the whole way
The typical research engagement ends with a dataset and a deck of tables. Ours are built around everything that happens before and after that handoff. Each engagement travels the same arc, and we stay on the line for all of it:
Advise
Before anything fields, we pressure-test the plan: study design, sampling strategy, question wording, and the sharper question of what you would need to see in order to act.
Collect
Rigorous fieldwork matched to the design, with quality controls and transparent documentation, whether that means a 50-state poll, a randomized experiment, or forty hours of interviews.
Analyze
Modeling that matches the design and honest uncertainty alongside every estimate. We report what the evidence supports and flag what it doesn't.
Deliver
Products people actually use and share: infographics, interactive online tools and data explorers, and primetime-ready reports built for boards, funders, journalists, and policymakers.
Implement
We stay at the table for the strategy: message testing, intervention design, audience targeting, and communications guidance, plus the next round of questions the results raise.
Common needs, research solutions
Every engagement is custom-built, but the needs that bring nonprofits and businesses to us tend to rhyme. A few patterns, and how we solve them:
→"We need to be the credible voice on our issue."
Recurring polling and citable public reports turn an organization's authority into a steady stream of original data that press, policymakers, and members reference by name. Tracking cadence beats one-off snapshots: it shows how opinion moves and gives you something fresh to publish all year.
Recent work: multi-year national tracking of public attitudes toward emerging technologies, published openly and cited in national press.
→"We need to know which message will actually work."
Randomized message experiments replace intuition with evidence: real audiences, competing frames, measured effects. Paired with segmentation, they tell you not just which message wins overall, but for whom, so budgets go where they move people.
Recent work: randomized framing experiments on a contested medical treatment for a national research initiative.
→"We're launching something new, or defending something contested."
Baseline studies and audience segmentation map the terrain before you commit: who supports you, who is movable, what the sticking points are, and which misconceptions need to be corrected rather than argued with. Interviews and focus groups put texture on skeptics the numbers can't fully explain.
Recent work: post-election analysis of a statewide ballot measure, connecting vote patterns to messaging lessons for the next cycle.
→"We need to show our program works."
Randomized and quasi-experimental evaluation designs produce the causal evidence that funders, boards, and journals take seriously, built with measurement that captures the outcomes that matter rather than the ones that are merely easy to count.
Recent work: outcome measurement and survey infrastructure for national health-focused nonprofits.
→"We have data. We need the story and the strategy."
Sometimes the data already exists and the gap is everything after: rigorous reanalysis, visualization, and translation into concrete guidance on audiences, framing, and next steps. We turn archives into assets.
Recent work: transforming a partner's multi-wave survey archive into public-facing reports, interactive graphics, and a communications playbook.
Want to work together?
We're in the business of finding answers. Let's talk about your questions.
Start a conversation