How do you avoid confirmation bias and cognitive dissonance in your lean startup experiments?

Confirmation bias can be so dangerous at an early stage. In the Lean Startup methodology the hardest but most important aspect to nail down is actively listening to the market and taking decisive action on what’s really being indicated, not what you want or hope is being indicated. This is how you build products/features/companies that are set up for scale and potential mass adoption.

El sesgo de confirmación es un fenómeno cognitivo en el que las personas tienden a buscar, interpretar y recordar la información de una manera que confirme sus creencias preexistentes, mientras descartan la evidencia que contradice esas creencias. Para prevenir el confirmation bias, se pueden tomar varias medidas: 1. Mantener una mente abierta 2. Buscar activamente evidencia contraria 3. Fomentar un ambiente de equipo crítico 4. Utilizar datos objetivos 5. Implementar pruebas de validación rigurosas En definitiva, hay que ser consciente de ello y buscar refutar nuestras creencias de forma activa.

Confirmation bias has been seen in the micromanagement bandwagon for over 30 years. There is a great belief that clinical interventions can reshape outcomes. Even with the understanding that outcomes are mostly shaped by social and other non-clinical drivers, the belief remains. Studies with poor controls, creative analysis, and few limitations add to the problem. The designers of US health care see overutilization as the enemy and studies are designed to confirm this. They do not see the negative where 40% of Americans have deficits of access where underutilization and inappropriate utilization are rampant.

Be honest with yourself, your team, and the data. Getting this right/wrong at this stage is often make it or break it. The more data points and subsequent methodology to asset it should be your framework at this early stage.

La disonancia cognitiva se refiere a la incomodidad mental que experimentan las personas cuando tienen creencias contradictorias, especialmente en situaciones de toma de decisiones. Un ejemplo en el contexto del lanzamiento de un nuevo producto sería creer firmemente en la innovación del producto, pero al mismo tiempo, enfrentarse a ventas iniciales bajas y críticas negativas. Para combatir este efecto, es importante contar con: • Objetivos claros • Comunicación transparente • Flexibilidad y adaptación • Capacitación y apoyo para el equipo Estos elementos ayudan a alinear las expectativas y reducir la incomodidad causada por creencias encontradas.

El confirmation bias te lleva a buscar y recordar solo la información que confirma tus creencias, sesgando los resultados. La cognitive dissonance provoca incomodidad al enfrentar datos contradictorios, lo que puede hacer que ignores evidencia negativa para evitar cambiar de estrategia. Para combatir ambos suelo aplicar alguno de estos consejos: • Diseñar experimentos rigurosos y objetivos • Fomenta el feedback de forma abierta • Realizar revisiones cruzadas entre personas del equipo • Automatizar y mostrar el análisis de forma transparente. Finalmente también hay que ajustar la estrategias basada en evidencias, e intentar huir de decisiones basadas en opiniones personales.

This is where advisors or coaches can really help at this stage, or a BoD if you have one. Listen to those commuted to your success but who aren’t as close to the project. And being aware of these biases can help not fall victim to them. Be intentional and self actualized.

The focus at CMS has been on micromanagement of cost and quality from above and outside of practice environments for over 30 years. Health care costs remain out of control and outcomes are not improving. The models did not consider alternatives. Metrics chosen were vague or easily gamed. And more and more we understand that outcomes are largely shaped outside of clinical interventions. Studies continued to focus on micromanagement despite marginal results and higher costs of delivery. Not an emphasis are studies of practice environments, delivery team members, and the consequences of the least experienced CMS was 5 for 52 in successful innovations costing billions. In any business operation, their kind of innovation would be terminated.

In my experience, the simplest method to avoid bias in experiments is to separate the teams that benefit from the outcome from the teams that run the experiments. For example, getting a UX Designer to test the mock-ups they have spent a long time on is challenging. They have a vested interest in the success and the feedback, though valued, is likely to be a somewhat challenging process for them. This applies doubly to founding teams.

To mitigate confirmation bias, use the following strategies: 1.Seek contradictory evidence by actively looking for data that contradicts your hypotheses to counteract confirmation bias. 2.Use blind testing methods in testing to reduce bias from researchers and participants. 3.Design experiments with stringent controls to objectively test hypotheses. 5.Iterative approach follows the build-measure-learn-loop to emphasize learning from actual data, minimizing personal biases. 6.Utilize insights from external advisors for unbiased perspectives. 7.Keep thorough records of all assumptions and decisions to ensure they are data-driven. 8.Be willing to pivot based on what the data suggests, overcoming any resistance due to cognitive dissonance.