We analysed multiple feature flag solutions and use multiple criteria to narrow down our candidates. We are a small, fully remote team.
Apptimize
Site: https://apptimize.com/?r=mwf
Support TS/JS: Yes
Support BE & FE: Yes
A/B Testing: It supports A/B testing
https://apptimize.com/docs/abexperiments.html
Experiments supports Jira, so Jira can be utilized as well as updated from experiments
Feature Flag Management:
Supports Feature Flag management where, a feature can be switched on or off if needed or not
https://apptimize.com/docs/featureflags/setup.html
Support for Analytics: Yes
https://apptimize.com/docs/abexperiments/results.html#results
Analytics are driven by result Goals, give the system possible solutions, and it analyses based on the goal expectations.
Segmentation and Filteration
Events & Funnel https://apptimize.com/docs/additional-features/events-and-funnels.html
Data Exporting https://apptimize.com/docs/additional-features/data-exporting.html
Price Plan: No stated price plan
Special Feature:
Very convenient for mobile applications.
It has proper documentation for feature flag setup, allocation and targeting
It supports cross platforms Client-Side - Mobile/Web SDKs (Android, iOS, JavaScript)
Server-Side SDKs (Python, Node.js)
Instant Updates https://apptimize.com/docs/instant-updates.html
Preview Groups https://apptimize.com/docs/additional-features/preview-groups.html
Pilot Users https://apptimize.com/docs/additional-features/pilot-users.html
Flagship
https://www.flagship.io/?r=prd-ffs
Support TS/JS: Yes
Support BE & FE: Yes
A/B Testing: A/B testing for agile teams
There is no much information about testing though
Feature Flag Management: Support Feature flags & Toggle.
It uses Decision API as an engine to Flagship which is available on AWS API Gateway
Support for Analytics: Yes
With Flagship, analytics can be integrated with Mixpanel https://docs.developers.flagship.io/docs/integrate-with-mixpanel and Segment https://docs.developers.flagship.io/docs/integrate-with-segmentcom
Price Plan: No stated price plan
https://www.flagship.io/pricing/
Special Feature: It support webhooks for custom integration based on a feature state and flag state https://docs.developers.flagship.io/docs/webhooks-integrations
Flagship supports Remote Control API https://docs.developers.flagship.io/reference/getting-started
Its supports Test in Production
Instant Updates
Smart Decision API
Kill Switch
Split
https://www.split.io/?r=prd-ffs
Support TS/JS: Yes
Support BE & FE: Yes
A/B Testing: Splits power A/B tests in both BE and FE. Turning every feature you develop into an A/B test.
https://www.split.io/solutions/a-b-testing/
Feature Flag Management: Split provides everything you need to create, target, and manage feature flags. It ties feature flags to your engineering and customer data.
https://www.split.io/product/feature-flags/
misbehaves. And we calculate the impact of your code on customer outcomes.
Support for Analytics: Since it has external integrations, test data and result are exported to analytic services like segment etc.
Price Plan: It has a free account, providing some unlimited services
Special Feature: A data hub for features like google analytics, segment, sentry, jira, datadog, sumo logic etc.
It supports continuous delivery
Test in production
Molasses
https://www.molasses.app/?r=prd-ffs
Support TS/JS: Yes
Support BE & FE: Yes
A/B Testing: A/B testing is supported
Feature Flag Management: Its supported
Support for Analytics: no much information on this
Price Plan: It has Free package, $50/month package and $200/month package. Read for details
Special Feature: Detailed audit logs to give details behind a feature change plus alert notification
Flagsmith
https://flagsmith.com/?r=prd-ffs
Support TS/JS: Yes
Support BE & FE: Yes
A/B Testing: It supports A/B testing
https://docs.flagsmith.com/advanced-use/ab-testing
Feature Flag Management: Flagsmith provides an all-in-one platform for developing, implementing, and managing your feature flags.
It also supports multi-variate flags
https://docs.flagsmith.com/basic-features/managing-features
Support for Analytics: Flag Analytics allow you to track how often individual Flags are evaluated within the Flagsmith SDK.
https://docs.flagsmith.com/advanced-use/flag-analytics
You can also integrate third party analytic integrations for comprehensive analysis
Price Plan: Detailed pricing plan
https://flagsmith.com/pricing/
Special Feature: Its can be used for crossplatforms
It has additional features like change requests, Edge API, Edge Proxy, Flag Management, Role based Access Control, Scheduled Flags, System Administration
Webhook integrations are supported for custom configuration
https://docs.flagsmith.com/integrations/webhook
There are more integrartions like Amplitude, AppDynamics, Datadog, Dynatrace, Heap Analyticd, New Relic, Rudderstack, Slack and Terraform.
It can also be deployed on Docker, Kubernetes, AWS, Google Cloud, Openshift etc
https://docs.flagsmith.com/deployment/overview
Optimizely
Support TS/JS: Yes, they have fullstack sdk
Support BE & FE: Yes
A/B Testing: A/B testing, Multi armed bandit and Stats acceleration, Mutual exception groups, Allowlisting
Feature Flag Management: Has online mgt system, supports graceful rollout and rollback, pause. Can separate experiments by different env. Adjust params of the experiment at runtime.
Support for Analytics: Has support of major analytics
Segment, amplitude, mixpanel
Price Plan: I found that online
https://splitbase.com/blog/optimizely-pricing
Looks expensive
Special Feature: Advanced work with target audience https://docs.developers.optimizely.com/experimentation/v3.1.0-full-stack/docs/target-audiences
Has internal stats engine that allows to observe and interprete results of experiments
https://docs.developers.optimizely.com/experimentation/v3.1.0-full-stack/docs/analyze-results
can export events and all related data
it’s possible to setup experiments both from code and mgt ui
provide sdk agent for microservices architecture https://docs.developers.optimizely.com/experimentation/v3.1.0-full-stack/docs/optimizely-agent
LaunchDarkly
Atlassian uses them.
Support TS/JS: Yep, anything
Support BE & FE: Yes
A/B Testing: A/B testing, feature flags, Multiarmed bandit. Custom algorithm experiments.
Feature Flag Management: https://docs.launchdarkly.com/home/flags
Online system for organizing and managing flags. Has all kinds of features, turn on/off, link to other flags in order to trigger, etc. provides debugging capabilities (streaming flags usage in realtime), custom settings per env
Support for Analytics: Surprisingly, but exporting data is comparatevily bleak
https://docs.launchdarkly.com/home/data-export
However, segment is supported, but for enterprise customer.
Price Plan: Looks affordable
https://launchdarkly.com/pricing/?r=prd-ffs
Pro plan is nice
Special Feature: Provides scheduling, a planned activation/deactivation. Has approvals flow, that can prevent unchecked experiments from being exposed.
Integration with Teams, Slack, it allows observing experiments mgmt. Api for triggerring an experiment.
DevCycle
Support TS/JS: No typescript , but has JS SDK
Support BE & FE: Both client side and server side SDK
A/B Testing: A/B testing, feature flags, multiarm bandit. Variable features
Feature Flag Management: Standard Dashboard with 4 patterns of feature flags (by Martin Flowler standard)
Support for Analytics: They have something internal, called feature reach. U can also create custom metrics in their dashboard.
Apart from that they store user data in edgeDB and insist, that we are free to do whaterver we wnat with it. At least i found integration with Zapier.
Price Plan: Seems reasonable, I don't expect it to be more than 50 dollars a month. Not to mention they have a free plan for a single dev.
Special Feature: Nice examples of integration with serverless
Custom targeting rules per each experiment.
Experiments can be divided by environments (3 from the start) Very flexible variation constructor that allows custom set of props and values per each experiment.
They have real-time system to track feature flags changes. Their sdk support it.
Integrations with github and jira, that show you what feature flags being used in the project.
CloudBees
Support TS/JS: Has both JS and TS
Support BE & FE: Yes, both FE and back-end
A/B Testing: Feature flags, A/B and feature releases. No complex scenarios.
Feature Flag Management: This is a platform based on Jenkins. So this tool not only for feature flags but for CI/CD as well. Visusally not friendly. https://docs.cloudbees.com/docs/cloudbees-feature-management/latest/feature-flags/about-feature-flags . Stop, freeze, approve - all those actions are availble.
Capable of dual way configuration: create flag from UI or declare it in yml file of your repo.
Support for Analytics: They have own metric collection and visualization system in cloudbees, but insist that we can send it anywhere
Price Plan: This is kinda tasty, they have free plan for teams like ours
https://www.cloudbees.com/c/feature-management-free-trial-sign-up
Special Feature: Integration with Jira, github, Teams and Slack
https://docs.cloudbees.com/docs/cloudbees-feature-management/latest/integration/teams
Have special mode for testing locall or during e2e tests
GrowthBook
Support TS/JS: Yes
Support BE & FE: Yes, backend and frontend is supported
A/B Testing: Feature flags, A/B testing
Feature Flag Management: Growthbook UI creates and manages the Features and experminents. Those items are metadata that can be accessible via API or synchronized by webhooks as a JSON file. https://app.growthbook.io/
https://docs.growthbook.io/overview
Support for Analytics: They have inhouse analysys engine, but rely on external data sources for that. Good thing that they support almost anything. Mixpanel, segment, name it.
https://docs.growthbook.io/guide/mixpanel
Price Plan: It’s open-source, but they have a cloud hosted version that is comparatively cheap https://www.growthbook.io/pricing
Special Feature: Their biggest benefit so far - is a possibility to setup a self hosted instance with 99% of the functionality from the enterprise version. This will require some mingling with mongo and docker though.
Statistics is part of the product out the box. They also have encryption capabilities for SDK endpoints in pro-version.
https://www.growthbook.io/compare/growthbook-vs-launchdarkly
