How Applied Behavioral Science Teams Create Impact

Behavioral Science teams are spawning throughout departments and industries, the growth of which I recently detailed. Many internal consulting or shared services groups take on projects to help weave behavioral science insights throughout their firm's products and services. Others have embedded themselves into specific product/service lines, working with more established groups such as Data Science, Marketing, and User Experience. These approaches yield many ways teams generate value. While the field is still broad and there is much to be defined, I have created a framework describing how I, working on a Behavioral Science team, have impacted my organization.

Almost all teams apply a few Behavioral Science methods (for more details on methodology, check out this article).

The first method is the application of existing Behavioral Science lessons to build a research-backed product or service. We can think of an application that leverages social norms (sharing the number of friends already using it) to encourage users to engage. The team would leverage the great body of academic research on how social proof can promote action.

The second is the development of interventions to change behavior. To envision this, think of a behavioral science consulting firm that experiments with multiple solutions, choosing the most effective; for example, a series of conversations for managers and employees to encourage these 1:1s to happen more often. This same example could be turned into a digital product by a team at a technology company. Leveraging established research insights and developing evidence would best serve an individual or group in developing a behavioral science team. Doing so will set them on the right path to quickly create value.

When establishing a Behavioral Science team, the employees must set a mission of what they hope to do and the types of projects chartered to accomplish it. The way they approach creating value will be heavily dependent on two factors.

1. The first is the composition of the team (more quantitative or qualitative specialists), and
2. The second is where they find themselves within the firm (shared services team, product aligned team, within a more established group such as data science).

Regardless of these factors, four categories of work spawn out of a matrix of two dimensions. Depending on the development status of the product or service and whether the research will directly support it or will serve more forward-looking exploration, projects will fit into one of these categories.

Each category comes with an estimated time to value creation and an inherited risk of whether evidence will be helpful to the team. Each also needs to have established trust with product teams and more competition to get one's position to be the accepted path forward.

Dimensions

The first dimension is temporal, whether the team will work on established products and services or look forward to future ones. Working on current solutions brings live data as one can immediately see how people are reacting and the successful parts compared to those that are not. Although, since the item's launch, the team will face the uphill battle of convincing stakeholders of the needed adjustments and, overall, less freedom of potential changes. Working on future solutions comes with fewer parameters around the interventions that can be deployed but at the cost of less opportunity for immediate impact.

The second dimension is whether the team will influence specific products/services or will serve to develop research eminence to embolden the underpinnings of the product or service. Working on products means getting research to support the development of new features, campaigns, or whatever the team needs to grow a successful solution. The opportunity for impact is highest here, but so is the chance that work will only serve confirmation bias. Alternatively, producing thought leadership gets ahead of creating products and services by helping the broader team understand the field they are operating in but at less opportunity for immediate impact.

Four Sections of Impact

Current/Connected: Improvements

Let's begin with the category most likely to create value and the hardest to pull off: Improvements. Behavioral Science teams can make product and service improvements in many ways; the challenge is solving the correct problems and communicating them so that decision-makers will select the suggested path over others. At its simplest, this involves assisting a product team by bringing helpful, relevant research. It is experimenting to understand the most likely intervention to create valuable behavioral change at its most extreme. Making improvements spans the entire development cycle, starting with understanding how a solution attempts to induce behavioral change by measuring if it does. When done right, generating improvements begins with understanding the root behavior the product/service team is trying to change, finding research to point towards the correct mechanism, and ensuring that it gets built so that people will want to use it.

Creating product or service improvements connects directly to business outcomes such as usage and retention, giving it the most significant opportunity to generate value. Similarly, empowered with the quantifiable value that their products/services create, marketing teams can leverage this to encourage further usage. For a product leader to act on it, the team must make a case more robust than those who suggest improvements in user experience and marketing. Thus, creating improvements is something all teams should strive to do while recognizing that it may take time to build the trust required by product leaders to act upon suggested modifications.

Research Examples: Academic literature reviews for solution ideas; experiments to determine the most effective intervention; studies to determine the effect of features; and pre-mortems to envision what might go wrong once launched.

Improvement Case Study: Zillow

Future/Connected: Planning

The following category can save significant money later in the development process when done right. Planning involves doing research that helps strategize where a product or service should go in future iterations. Behavioral Science teams can find a place on a product roadmap by generating research that will help answer questions that are likely to arise. Sometimes this means reviewing current features to determine what improvements could occur in the next version. In other instances, this could be embedding themselves with teams at the early stages of entirely new services.

Planning, when done right, prevents rework, which provides an instant case for why teams should do it. Although, behavioral science is undoubtedly not the only group that will do these types of work. Again, it will require building trust with product teams to adhere to the recommendations generated by any insights. Successful planning should connect to essential metrics such that insights are likely to maximize whatever the product team desires most. Similarly, it will try to factor in many different perspectives to avoid bias further down the development process. Ultimately, planning can yield tremendous value when done right and listened to but quickly becomes useless if it misses the stakeholders' expectations or doesn't answer relevant questions.

Research Examples: Academic literature review for related new product features; qualitative studies to determine what variables of an environment users find most important (conjoint analysis); examinations of current features to determine how future ones could better utilize habit formation insights.

Planning Case Study: Uber

Current/Positioning: Context Setting

Shifting to positioning impact, research which helps set context around the environment in which a product or service takes place can be extremely helpful in ensuring the value created by the solution occurs. This type of work revolves around defining how the behavioral change happening due to a solution connects to more considerable business outcomes. Suppose a product team establishes that their solution effectively engages workers in more exercise. In that case, this project category will have the goal of connecting this effect to the value created through physical fitness. Ideally, this type of work ultimately justifies why an individual should use the product or service and can be surfaced within it to encourage continual engagement.

The significance that context-setting work pursues connects to the value of the feature overall generated. A behavioral science team should engage in this work to validate the claim that a product or solution has significant value when used. Context-setting work also encourages the organization's image as a thought leader in its operation space. The most important risk of this type of work occurs when a team defines the context in which a feature operates, only to find no change to the business outcome through usage. This situation is essential to discover but requires the product leaders to act upon it, setting up the potential for conflict.

Research Examples: Partnering with academic institutions to connect product usage to important outcomes; quantitative studies to ensure that metrics within the service as meaningful and correct.

Contextual Case Study: BetterUp

Future/Positioning: Orientation

The final category, which is hardest to connect to immediate impact while also having tremendous potential to improve a product line or company drastically, is direction establishing work. Behavioral science teams, especially those bonded with the academic and cutting-edge research world, are uniquely positioned to bring in new insights that can drastically alter the strategic path of a product or service. This research intends to support the product vision while pushing it into new areas to maintain something fresh and valuable.

Investing in direction research means the group is willing to explore what is out there to pursue, ideally aligning their product with future customer needs, emerging markets, and new technologies. This type of research generally takes longer and commonly does not always yield helpful insights. Although, when successful, it can help develop a pipeline of science-backed features and demonstrate leadership in the product team's space. Teams that venture into this space need to be supported by a team willing to wait for results.

Research Examples: Qualitative research to understand users' habits, especially behaviors they are trying to change (diary study); Academic literature reviews to determine emerging behaviors that might alter with a product or service.

Direction Case Study: MoneyMap