Is there any way to reduce the confidence of an experiments to get results faster? For example, our marketing folks said that if we can get an A/B test with 75% confidence, that's enough for them to try more tests instead of waiting for a 90 or 95% result.

Hello @fresh-football-47124 👋 I didn't setup experiments for a while. Now I'm setting up a new one, but it seems that in the meantime you removed the ability to set a JSON for each variation (on cloud version). Can you confirm? If so, is there a way to get a set of parameters for the assignment variation, without having to implement it in the client code? Thank you

Hello, We are implementing growthbook with nuxtjs. We encounter problem of how to switch between variations. It seems that

growthbook.forceVariation

does not work. Is there some other way to force variation? When initializing the feature information is:

{
  "value": true,
  "on": true,
  "off": false,
  "source": "experiment",
  "ruleId": "",
  "experiment": {
    "variations": [
      true,
      false
    ],
    "key": "webtest-feature",
    "weights": [
      0.5,
      0.5
    ],
    "hashAttribute": "session_id"
  },
  "experimentResult": {
    "inExperiment": true,
    "variationId": 0,
    "value": true,
    "hashAttribute": "session_id",
    "hashValue": "45274132-7196-4e23-8f8c-ceece9fc9d16"
  }
}

After

growthbook.forceVariation('webtest-feature', 1)

it seems to be:

{
  "value": true,
  "on": true,
  "off": false,
  "source": "defaultValue",
  "ruleId": ""
}

Would love some help here. We have the same component in two places but they see two different versions of the A/B test. This obviously makes the measurements invalid for those cases but it's weird because we use the React hook. Could it be related to https://growthbookusers.slack.com/archives/C01T6PKD9C3/p1649706438476469?

hello everyone, im try to run in k8s. but i got error this

$ wsrun -p '*-end' -m start
back-end
 | $ node dist/server.js
front-end
 | $ next start
 | ready - started server on 0.0.0.0:3000, url: <http://localhost:3000>
 | info  - Using webpack 5. Reason: no next.config.js <https://nextjs.org/docs/messages/webpack5>
back-end
 | Killed
 | error Command failed with exit code 137.
 | info Visit <https://yarnpkg.com/en/docs/cli/run> for documentation about this command.
 | `yarn start` failed with exit code 137

Is it a known issue that the Chrome Extension doesn't work with PWA's?

Hello! Is it possible somehow create relative metric? By saying "relative", I mean something like this: I want for example OTS (open-to-search) metric. Right now I have two binomial metrics - open and search (yes/no, number of users who opened something and then number of users who find something and it's all with me experiment feature) then divide them, and now I want to check this metric in my experiment; test group - 45% OTS and control 40% OTS (and its significant). I hope I was able to explain

Hello everyone, is it possible to combine rollout rule with experiment rule? There could be a potential scenario, i want to do feature experiment within a small group of random users and figure out the winner, then i choose to rollout the feature to all the users with the winning variant

Hi all, any advice on how to configure an experiment to analyze only logged-in users? I have defined a metric to be for logged-in users only but the query for the experiment analysis looks at anonymous users only. What am I missing here?

Hello everyone 👋 Appreciate a bit of help. I am currently using

@growthbook/growthbook-react

with Next.js; current issue is that 👉 How can I get an experiment inside

getServerSideProps

func and track it to segment too (

window.analytics

) ?

Hi GB team 🙂 I was wondering how to interpret the risk thresholds in the metrics advanced options: • If my goal is to increase a particular metric, is it the risk of that metric decreasing by less than 0.25, between 0.25 and 1.25% and more than 1.25%? • How can I know whats the best strategy regarding risk management, in other words, to which numbers I have to set the thresholds? fyi @stocky-nest-87172

Hi all, is there a way to limit team member privileges on modifying/disabling production-enabled feature flags? I'd like to invite our whole organization to GrowthBook so that anyone can view and comment on experiments. However, I'm a bit worried that with increasing number of users we have an increasing risk of unintentional edits in feature flags with immediate effects in our service. It seems that the current roles only concern Experiments and general admin stuff - does any of the roles also restrict Feature editing?

Hey everyone, is there a way to purge the GrowthBook CDN and force a refresh? There’s a lot more context to this problem, but the summary is that we changed an attribute from a

string

->

number

and back and it’s still coming through as a

number

in the “Forced Value” rule. (i.e. our array for the “$in” operator is coming through as [1, 2, 3, 4] instead of [“1”, “2", “3”, “4"]) So all of our forced values aren’t applying for this rules since it’s trying to match the

string

we’re providing with a

number

.

Hey everyone. Is there a way to run an A/A test using a boolean feature?

Hi, is it possible to fire a webhook upon experiment creation rather than feature definition?

is it possible to make code highlighted in places like here too? sometimes I make typos in metric's script, and highlight could help in cases like this

👋 Hi everyone! I have a question about your product, who should I address it too

I see really strange issue on the same data..

👋 Hi everyone! is there an API endpoint to enable an environment for a feature?

Hi! I'm wondering about the implementation of the SSR single-page app. How does the analysis system work if we track the A/B experiment view once in server-side render and once when the app hydrates on client-side? Both saved values would have the same session_id as a unique identifier but the timestamp will vary. I know that the

trackingCallback

fires only once per experiment view but our growthbook instance in server render and client-side do not communicate between each other and the view event is sent once on both. Do we need to drop another one when making a query to the analysis, or can we leave it as is and analysis recognizes those experiment viewed events as coming from the single session? Or how have you worked around this? The feature we have affects on server-side code and client-side code so it will be triggered on both sides for the same user.

Hi all, is this a correct use of namespace? I want the experiment to run for 10% of all traffic. 5% per variation. Is it correct to put .05 for each? Or is that relative to the coverage I set in the namespace?

It's relative to the namespace coverage. So you want to put .5 for each variation instead of .05

Hey growthbook team! I really love the concept of the User Id Join Table that you can optionally configure in each source. However, I have added SQL there that stitches user_id and anonymous_id together, but it is never actually part of any queries in my experiments. When I configure analysis in an experiment to use user_id instead of anonymous_id as User Id column, is when I expect this user id join table to be included in the query, is this the case? I remember seeing it part of my queries once, but haven’t been able to replicate it since

Morning everyone We are thinking we taking the Cloud version of Growthbook but i couldn't find anywere how much traffic (Api calls) it can handle. Does it scale ? Thank you for any insights

Hello everyone, we currently have some AB tests done through GrowtBook but it seems we only get a sample of the data. E.g., we have a page that gets more 2 000 unique page views per week (based on Google Analytics and other sources) but Growthbook shows that we only have about 400 people in the AB test running on the page. Our GrowthBook is connected to GA and even metrics seem to work but the data is different. Any ideas why this could be? :)

Hi folks, Huge thanks to Growthbook community for such a stylish & functional tool. I was wondering if there’s a way to add a metric that should be cohort-level, not user-level (e.g. overall CTR across the cohort or average (across the cohort) position in the ranked list where conversion happens); i.e. aggregation should not happen before summing values for all users.

Hey team! We’re self hosting growthbook by running the image on kubernetes with a separate Mongodb hosted on GCP. Everything works flawlessly except for one thing which is - every time we restart the docker container, all images uploaded to experiments get greyed out, and there’s only an image icon left. I will upload a photo so you can see how it looks - any ideas on what goes wrong?

I keep getting locked out of my account with this error`You must verify your email address before using GrowthBook`

do you guys have a comparison chart of any sort depicting you against LaunchDarkly and VWO FullStack by any chance?