Last updated November 15, 2026
In Stephen Wendel's 2020 global survey of applied behavioral science teams, covering 324 of the 597 organizations known to have such a team, the median team had four people and had run five experiments in the previous year, about one every ten weeks. Just over half of the organizations on that list (53%) were consulting or contracting firms rather than in-house teams.
Build an in-house behavioral science team when behavioral questions come up in most planning cycles and leaders will act on test results. Hire a behavioral science consultancy when demand comes in bursts or the work needs methods or an outside view your team lacks. Use a hybrid model when you need a permanent internal capability plus occasional specialist depth.
At The Decision Lab, we support all four models described below, from defined external projects to designing and staffing a permanent behavioral science team for one of North America's largest insurers. This guide is for product, customer experience, HR and strategy leaders deciding how to add behavioral science capability without ending up with an underused internal team.
Four ways to get behavioral science expertise
- Hire an external consultancy for a defined project. This suits organizations with one or two high-stakes behavioral problems a year and no ongoing pipeline.
- Build an in-house team. This suits organizations where behavioral questions come up every month across products, customer journeys or employee programs.
- Combine a small in-house core with external specialists. This suits steady demand with occasional needs for skills such as large field experiments or machine learning analysis.
- Hire a consultancy to build your team. This suits organizations that have decided to build but have no one internally who has led a behavioral team before.
Build vs. hire at a glance
| Factor | Lean toward building in-house | Lean toward hiring a consultancy |
|---|---|---|
| Capability depth | Needs center on one or two skills, such as digital experimentation | Needs span diagnosis, design, experimentation and implementation |
| Project volume | Behavioral questions arise in most planning cycles | One or two high-stakes problems a year |
| Cost structure | Steady demand justifies a fixed cost | Demand is uneven or still unproven |
| Knowledge transfer | Methods must live inside the organization long term | You need results now and can contract for a full handover |
| Outside perspective | The problem is well understood internally | The problem is politically sensitive or new to the organization |
If most of your answers fall in the left column but one or two land on the right, a hybrid model usually fits best.
Build vs. hire: the five factors that decide it
1. Capability depth
Applied behavioral science covers several distinct skills: diagnosing why people act as they do, designing interventions, testing them in controlled experiments, and building the winning version into a product or process. A four-person team, the median in Wendel's 2020 survey, rarely holds all of them at depth. In the same survey, nearly a third of internal teams sat inside data science and a similar share inside product teams, with standalone behavioral teams relatively rare.
If your needs cluster around one skill, such as experimentation in a digital product, a small internal team can cover it well. If you need the full range, a consultancy gives you access to specialists you could not justify hiring full time.
2. Project volume
An in-house team pays off when there is enough work to keep it busy on problems that matter. If behavioral questions come up in most planning cycles across several teams, you have a pipeline. If they come up once or twice a year, you have a project.
Volume also decides whether a team's experience compounds. The low experiment count in Wendel's 2020 survey may reflect barriers to implementation and measurement more than a shortage of ideas. About 2 in 5 respondents named getting interventions implemented as their biggest challenge, and a similar share named measuring impact, compared with about 1 in 7 who named generating ideas.
3. Cost structure
An in-house team is a fixed cost. You pay salaries and recruiting costs, plus the experimentation tools and data access the team needs, whether or not there is a high-value problem that quarter. A consultancy is a variable cost tied to defined projects, at a higher rate per day but with no idle time and no hiring risk.
The comparison that matters is cost per decision changed. An internal team that spends its first year building credibility and data access may change fewer decisions than a consultancy engaged for two well-chosen projects. Once an internal team has a steady pipeline and leadership support, its cost per tested decision will often fall below the cost of repeatedly commissioning comparable external projects.
Building also costs leadership attention. In early results from Wendel's 2023 survey of behavioral science practitioners, the two most common challenges were measuring the impact of their work, named by about 2 in 3 respondents (65%), and making the case for behavioral science to senior leaders, named by a similar share (63%).
4. Knowledge transfer
Hiring externally carries the risk that the expertise leaves when the engagement ends. Ask any consultancy what your team will own at the end: the methods, the data and analysis, the intervention materials, and the ability to rerun the test.
In our AI adoption work with a Fortune 500 HR technology company, each of the eight programs we piloted came with the protocols and guides the client's own teams needed to run it themselves. Those programs are now scaling to the company's full 20,000-person workforce.
Building in-house keeps knowledge inside the organization, but only if the team documents what it learns. A behavioral team that loses its lead can lose most of its institutional memory with them.
5. Need for an outside perspective
Some problems are hard to see from inside. Colin Camerer, George Loewenstein and Martin Weber named this the curse of knowledge in a 1989 paper in the Journal of Political Economy: once people know something, they struggle to predict how someone without that knowledge will judge or act. An internal team that knows the product, its jargon and its history is exposed to the same bias when it imagines a first-time customer.
An outside team also brings patterns from other industries, and can deliver findings that are politically difficult for an internal team to raise, such as evidence that a flagship program does not change behavior.
What a behavioral science or choice architecture consultancy adds beyond an in-house insights team
Many organizations already have an insights team running customer research and ask whether it can absorb behavioral work. The two overlap, but they answer different questions. An insights team usually measures what customers say and think: their needs, attitudes, satisfaction and preferences. Behavioral science aims to change what people do. Much of that work is choice architecture, the design of how options are presented to people (defaults, order, framing and the effort each choice takes), and it has to be tested against what people actually do.
| Typical in-house insights team | Behavioral science consultancy |
|---|---|
| Measures attitudes, needs and stated preferences | Diagnoses why people do not act on what they intend |
| Surveys, interviews, focus groups and tracking studies | Field experiments and A/B tests with a control group |
| Recommends what to change | Designs and tests the specific change |
| Reports findings to decision-makers | Stays involved until the tested version is built and running |
The strongest setups connect the two. The insights team finds where customers drop off or hesitate, and the behavioral team designs and tests what to change at those points. For a fuller comparison of the disciplines, see our guide to behavioral science vs. UX research vs. market research vs. management consulting.
Problems that benefit from an outside perspective
- Decisions that cut across departments, where no internal team has the mandate to redesign the whole customer or employee journey.
- Evaluations of a program leadership already believes in, where an internal team has a stake in the answer.
- First-time behavioral work, when nobody internally can judge whether a proposed experiment is well designed.
- Large or unusual studies, such as analyzing thousands of hours of call recordings or running a field trial across several markets, that would take an internal team months to set up.
- Ethical review of choice architecture, to check that defaults and framing serve customers as well as the business. In Wendel's 2020 survey, about 2 in 5 responding teams (41%) said virtually none of their users knew about the behavioral interventions applied to them.
How the hybrid model works in practice
One of North America's largest insurers asked us to help it build behavioral science into its digital and customer experience strategy as a permanent capability, instead of buying it project by project. We based the team's design on one-on-one interviews with 25 operational stakeholders and a review of the insurer's internal processes and products, with input from advisors who had built behavioral teams at the White House and Google. The result was an organizational blueprint and hiring profiles that combined behavioral science with digital expertise, and the team was hired.
We also ran a first project to prove the team's value. From a ranked list of candidate pilots, we chose agent sales calls, analyzed more than 170,000 hours of calls with process mapping and machine learning to find where conversations went wrong, and piloted script changes and agent training against a control group. Sales rose 12% in the pilot group compared with the control group, with a projected increase of $35 million in annual revenue. Read the case study on how an insurer built an in-house behavioral science team and projected a $35 million annual revenue gain for the full details.
After launch, the usual pattern is for the in-house team to own the steady pipeline and bring in outside specialists for large studies, new methods or problems where its own view is too close.
How we work with in-house teams at The Decision Lab
We are an applied behavioral science consultancy with more than 60 published case studies. With organizations that have, or want, their own behavioral capability, we usually work in two ways: adding specialist depth on projects an in-house team cannot staff, such as large-scale analysis or controlled field tests, and helping organizations design, hire and launch a team of their own. We also run skill-building work. For the Open Society Foundations' Strategy Unit, we identified the skills behind effective strategy and built skill-building resources to connect strategy and implementation.
If you have already decided to hire externally, our guide to choosing a behavioral science consulting firm covers what to ask on a shortlist. We are probably not the right partner if you need a one-off survey or focus group study, or recommendations you do not plan to test. Your own insights team or a market research agency will usually serve you better there.
Frequently asked questions
How many people do you need to start an in-house behavioral science team?
Many teams start small. The median applied behavioral science team in Stephen Wendel's 2020 survey had four people, and an OECD survey of government teams found almost 6 in 10 had fewer than five people applying behavioral science. Size matters less than access to data and a senior sponsor who will act on results.
Is an in-house team cheaper than a behavioral science consultancy?
That depends on volume. An in-house team is a fixed cost, so it becomes cheaper per project once there is a steady pipeline of work. With uneven or unproven demand, a consultancy usually costs less, because you pay only for defined projects and carry no recruiting risk or idle capacity.
Can our insights team take on behavioral science work?
Insights teams can take on part of it. They already run research and understand customers, and many members can learn behavioral diagnosis. The usual gaps are designing interventions, testing them against a comparison group in a live setting, and the authority to change the product or process being tested. Joint projects with specialists can close some of that.
What should a consultancy leave behind when it works with an in-house team?
At minimum, the methods used, the data and analysis, the intervention materials, and enough documentation for your team to rerun or extend the test without outside help. Write this into the contract. In our engagements, we aim to hand over each intervention with the protocols and guides the client's team needs to run it.
When should a company hire a consultancy to set up its behavioral science team?
When leadership has decided to build but nobody internally has led a behavioral team before. An outside firm can define the team's mandate, hiring profiles and first projects from evidence about the organization, then run an early project that shows results before the team is fully staffed.
Planning your behavioral science capability
If you are weighing whether to build a team or bring in outside help, talk with our behavioral science consulting team about assessing your current pipeline and choosing a first pilot worth running.

