Pros y Contras de Statsig

¡Statsig ha superado mis expectativas con creces! Nuestra empresa nunca había tenido ningún tipo de prueba antes, y cuando estábamos buscando plataformas, era difícil saber lo que necesitábamos. Pero el equipo de StatSig estuvo muy dispuesto a trabajar con nosotros en el precio del contrato y a ayudarnos con la configuración e implementación. Son muy receptivos a cualquier pregunta que tengamos, ya sea una pregunta de implementación técnica o una pregunta sobre cómo explicar algo complicado a un interesado. La herramienta en sí ha sido excelente: es fácil de usar, lo que realmente ha ayudado con la adopción entre el equipo y a comenzar a construir más una cultura de pruebas en nuestra empresa. Su equipo también es muy receptivo a las solicitudes de funciones y realmente implementa los comentarios en su hoja de ruta; muchas empresas dicen que harán esto durante el proceso de ventas, pero luego nunca lo cumplen. ¡Recomendaría StatSig para cualquier empresa que comience con la experimentación y las pruebas! Reseña recopilada por y alojada en G2.com.

La única desventaja que puedo pensar es que es solo una plataforma de experimentación, así que tengo que usarla por separado de nuestras otras herramientas de análisis de productos. ¡Pero eso no ha sido un verdadero problema! Reseña recopilada por y alojada en G2.com.

+ Funciona y es confiable.

+ Está bien documentado.

+ Muchos adaptadores (React/Vanilla JS, HTML, etc...)

+ La función de experimentos es muy buena.

+ La integración es muy fácil. Reseña recopilada por y alojada en G2.com.

- La documentación del SDK es incorrecta (dice "env" en lugar de StatSig env) - causó un error de nuestro lado.

- No se puede agregar descripción a las puertas de características - cuando una empresa crece y trabaja en múltiples cosas, los desarrolladores podrían no saber en qué están trabajando otras personas.

- No se pueden eliminar en masa las puertas de características - Debes entrar en las puertas una por una y eliminarlas.

- El SDK no admite la inicialización perezosa ni la actualización. Reseña recopilada por y alojada en G2.com.

It is packed full of advanced features and they keep adding new features while being very competitively priced.

It's easy to use and implement. I'm a big fan of their experiment results visualisations, everyone in our organisation can understand experiment results.

You get useful answers from customer support and not some generic scripts that I usually see from big vendors. Reseña recopilada por y alojada en G2.com.

You have no dedicated support for the pro-tier subscription but there are skilful Statsig people on Slack that you can ask questions and a useful support bot.

When asking more complex questions it might take a couple of days before you get a reply. When you factor in multiple questions back and forth with support + time differences between EU and US it may take a few weeks to get to the bottom of questions. Reseña recopilada por y alojada en G2.com.

As a data scientist, Statsig makes it incredibly easy to handle everything from metric breakdowns to performance tracking and experiment setup. I'm truly impressed by how it allows me to analyze all experiments in one place, with just the click of a button.

What's even more helpful is that Statsig doesn't just support setting up and analyzing experiments—it also lets you import and analyze experiments run outside the platform. This feature is a game-changer, making it seamless to get insights from experiments, no matter where they were conducted. Reseña recopilada por y alojada en G2.com.

The UI could be improved. For first-time users, it can be a bit challenging to understand why certain actions aren't working as expected. For example, some buttons are greyed out on certain screens, and it's not always clear how to proceed. It feels a bit like a trial-and-error process. Reseña recopilada por y alojada en G2.com.

I've used other experimentation platforms before from Meta's "Deltoid" platform to one I built myself in python... Statsig far surpasses the other platforms out there in terms of both ease of use and flexibility. It has been a game changer for experimentation across my company. The whole process is now seamless, from adding new metric sources & metrics, setting up scorecards for new experiments, and diving into more detailed analysis with things like days-since-first-exposure view & metric results split by user dimensions. Reseña recopilada por y alojada en G2.com.

There are a lot of options to pick from for metric types, experiment settings, etc. This provides a TON of flexibility in how you analyze experiments, which is great. But the number of choices can be a bit daunting when you are newly onboarding! So definitely ask questions to the Statsig support team, and they will promptly give you advice to help you with any choice you are making. Reseña recopilada por y alojada en G2.com.

Easy to set up experiments

Easy to view results (does not require manual data pull)

Great data visualizations - confidence intervals are so often misunderstood, having them visualized is wonderful. Reseña recopilada por y alojada en G2.com.

Cannot use Layers or Feature Gates with "Autotune" feature - makes it unusable for our team as we use these for safe roll outs.

Alerting on metrics - would like more options. E.g. metrics fluctuate a ton at the start of an experiment when traffic is low, so I don't want to be notified about that. I want to set alerting on experiments so that "if X core metric drops Y percent with confidence, send an email to this alias" Reseña recopilada por y alojada en G2.com.

As a product manager, I used statsig to track the performance of my website. It was really easy create various charts and funnel data with minimum steps. Without much technical knowledge I was able to track various events and get some insight into how our producr is performimng. Reseña recopilada por y alojada en G2.com.

With some guidance, I was able to set up charts quickly, But someone who hasn't worked with lots of data before, it's somewhat tricky. Reseña recopilada por y alojada en G2.com.

The flexibility it allows for custom rollouts. Layers have been a great feature used for more complex ml model rollout solutions.

In addition simple binary feature flags are a great way of getting things off the ground Reseña recopilada por y alojada en G2.com.

Have noticed some weird behaviours when the connection using a statsig client fails Reseña recopilada por y alojada en G2.com.

Statsig provides a platform for the entire lifecycle of an experiment. The clear distinction between different concepts like events and metrics enables teams to learn and adopt the industry-leading ways of running experiments. It has allowed my team to start experimenting within a month. Reseña recopilada por y alojada en G2.com.

The Statsig UI can be hard to navigate in the beginning. It takes a while to understand how to find individual features but they are getting better at surfacing them. Reseña recopilada por y alojada en G2.com.

StatSig has built the implemented the functionality that I enforces good Feature Flagging and Experimentation features that are often hard to get right. The ability to create experimentation layers, parameter stores, automatic metric calculations for lift are incredibly important and designed well. Unlike other feature flagging platforms that couples together Feature Flagging with Experimentation, StatSig does a great job of separating the two concepts. Their customer support through their public Slack is incredible, they're super responsive and provide very helpful answers to get us on board. Reseña recopilada por y alojada en G2.com.

There are Parameter Store and Dynamic Config, there should really only be one. Reseña recopilada por y alojada en G2.com.