Skip to main content

Product Experimentation

Amplitude Experiment is our product that allows teams to remotely modify and configure the product experiences for unique audiences.

  • 116 Topics
  • 196 Replies
116 Topics
F
Novice
Florent VilmartNovice
asked in Product Experimentation
Exposure tracking through segment

Hi there!For Amplitude experiments, it is recommended to emit an exposure event.We are using segment.io and are wondering how we can track those events properly.Whenever we call `track` it ends up being an event with name `$exposure`.Segment doesn’t seem to wanna pass `event_type` to the destination.The alternative is to use the `identify` call, but I wonder if there are side effects compared to using the $exposure call.

4717
F
Novice
V
Vania AlyaNew Member
asked in Product Experimentation
Error Message when Saving Experiment

Hi amplitude team, I wanted to raise an error message that I encountered when saving changes to my experiment. If i may know what triggers this error? Because this happened to a couple of experiments with the same error message and we don’t use any value ($1,$2,$3) that was stated in the error message. Thank you!   

1472
Y
Team Member
O
Ossi PesonenNew Member
asked in Product Experimentation
Amplitude Experiment device type for iPhone 14 not working

We are trying to set up A/B testing for anonymous users that is targeted specifically for iPhone users. Amplitude, however, seems to be unable to detect iPhones for the variant as per our testing. For reasons unknown the variant is not returned in the request, because Amplitude does not recognize iPhone via the user agent.We are using the JavaScript SDK and setting it up as per documentation, and I’ve done some debugging to see what type of user agent string is being delivered via network requests from a phone:Granted, this is from an iPhone simulator, but based on the information I gathered from the web, it is exactly as a real device, which also do not work. We had QA engineers try to enable this experiment as well with real iPhones on our website and we couldn’t get it to work.Here’s a snippet of the JS code:const amplitudeExperiment = window.experiment;if (amplitudeExperiment) { var div = document.getElementById('variant-info'); amplitudeExperiment.fetch().then(function() { co

4153
Esther Trapadoux
Community Manager
A
anandraprincNew Member
asked in Product Experimentation
Unable to fetch new file from GCS Bucket

I am new to Amplitude, I have configured a source to GCS for event data import. It imported first time when I have configured, but later whenI am pushing another files to GCP bucket with same extension and format it is not reading that file.

3147
M
N
Novita OliveraNew Member
asked in Product Experimentation
Difference between Exposure event with blue font and green font

I saw that sometimes my exposure event is in blue one and sometimes it is in green font. I read that this is due to the event that is outside the session.Can I interpret that blue exposure is for front end experiment and green exposure is for back end experiment? What conditions that make the exposure event is outside the session?Reference: https://github.com/segmentio/analytics-react-native/issues/654 

4542
Esther Trapadoux
Community Manager
N
Novita OliveraNew Member
asked in Product Experimentation
Best practices for using custom exposure in Amplitude experiments

Our aim is to utilize custom exposure in Amplitude experiments to make sure the users who see the changes is expected as the definition that we want to target. But we have observed a problem of receiving some exposures without any assignment, even after a few weeks of implementation. Is there any suggestion or guidelines of best practices for using custom exposure effectively in Amplitude?https://analytics.amplitude.com/tiketcom/chart/new/denicz1 

6281
Jeremie Gluckman
Team Member
I
Novice
invictusNovice
asked in Product Experimentation
Exposure Events Tracking

Hi Amplitude CommunityI am trying to launch my first experiment and trying to track exposure eventHowever the docs seem to not have been updated.What is the correct code to track exposure event.I am using this amplitude.track( "$exposure", (Map<String, ? extends Object>) new Properties() .put("exposure.getFlagKey()",exposure.getVariant()));} but I feel this is not the right wayCould someone please help me out?  

7687
J
Novice
L
Helpful Hero
liukaichiHelpful Hero
asked in Product Experimentation
Flag-Dependent Experiment Allocation

I have a question regarding experiment allocations.We have two flags that we want to be dependent on each other. For example "For users who are shown this, give them another experiment".For example using two flags, "show_tabs" and "show_new_shiny_tab". I would like to configure an experiment that is "For users who see the tabs, perform an A/B Test to users to test a new shiny tab". So to achieve behavior, we select on the "show_new_shiny_tab" experiment the following: "Users where show_tabs = true/false".  Will this work as expected? From what we see, there is room for doubt, but overall it looks like it is working. Is there any concerns with this approach? 

2611
Jeremie Gluckman
Team Member
J
jacoob_mNew Member
asked in Product Experimentation
Flag management by API.

Hi,I use a flag to enable user features in my app. I add user ids to the segments by UI.I did it according to the doc: Working with feature flags and feature rolloutsBut it is inconvenient with more users.I would like to do it by API. I got acquainted with Experiment customers: New endpoints available for the Management API but I did not find an answer there on how to manage the list of users belonging to the segment. I can pull a segment for a flag but I can't change the segment.Can I do it with the current state of the API? Maybe there are some other ways to enable/disable the flag for users via the API.

2861
Jeremie Gluckman
Team Member
F
Fran6New Member
asked in Product Experimentation
Cohort with specific behaviour every month

I would like to create a cohorte of people who performed a certain number of actions per month (like listen to < 20 songs in each month last year). But what if someone listen to 20 songsa month and then 21 in the other one ? In that case I would like to take him out of my cohorte and only keep the people who performed less than 20 songs every month during a year. Is it possible to get these people in a cohorte ?  Thanks in advance

2211
Saish Redkar
Expert
0
00rladkfk00New Member
asked in Product Experimentation
An error occurred while embedding Amplitude data into Google Sheets

Hello, community membersI have a question about integrating with Google Sheets.I followed the documentation on the Amplitude blog and tried to integrate with Google Sheets.I followed the instructions in the blog, but when I pressed the "Add" button, I got the error message and it didn't go any further.(* message -> ["Add Chart Failed: <html><title>500: Internal Server Error</title><body>500: Internal Server Error</body></html> ) I have searched within the community for information on this, but have not been able to find any.Does anyone know about this case. Any help would be great for me.Thank you 😀 

3472
Jeremie Gluckman
Team Member
A
aprbirdkimNew Member
asked in Product Experimentation
Can I delete my account and sign-up again?

Hi, I created my personal account with an organisation, and I realised that I have to re sign-up with a business account. Would it be okay to delete my current account and re-create it with the same organisation? Thank you :) 

5934
belinda.chiu
Team Member
Anthony
Peer Moderator
AnthonyPeer Moderator
asked in Product Experimentation
Android device blocked from sending events to Amplitude

I think my specific Android device, which has our Android app with the Amplitude Android SDK, is being blocked from sending any events to Amplitude. How can I check that? 

4594
Anthony
Peer Moderator
T
tcaprioNew Member
asked in Product Experimentation
Holdout Group option missing from UI

Why don’t I see an option to set up a holdout group in my account? https://help.amplitude.com/hc/en-us/articles/13508918823835

2302
T
I
Novice
IanNovice
posted in Product Experimentation
Snowflake data source not shown in Events as source

I added a full sync with valid SQL query from Snowflake, but the new data source cannot be found in Events, and cannot be loaded into Amplitude.

3085
Jeremie Gluckman
Team Member
N
neoNew Member
posted in Product Experimentation
HTTP V2 API "Connection reset by peer" error

There is no problem with the event sent to Braze (have the same structure), but the “Connection reset by peer” error occurs only in the event sent to Amplitude. (Occurs in about 5%) Since we are not using a separate proxy, there is no path for the error to occur on our side. What could be the cause? 

2332
Jeremie Gluckman
Team Member
R
reyanalystNew Member
posted in Product Experimentation
Snowflake Source not showing up in Event sources

This is a follow up to my concern in the link below:  I mentioned that Time-based ingestion for Snowflake integration is not working, but how about for the full-sync ingestion? Still does not work. Upon creating an event, you can click on the event to connect it to existing sources, in my case, I have 4 test snowflake sources. However, it is not showing. The number of sources say that I have 0, when I have 4. I actually tried to click add source to create another hoping it would attach the source to the event, but it would not. 

1931
Amplitude Admin
Community Manager
R
reyanalystNew Member
posted in Product Experimentation
Snowflake as Source is not working

I’ve tested 4 scenarios of different configurations to set Snowflake as source and all of them failed:Event + Time Based Configuration - successful creation of source BUT will have a Status of “No Data” even after performing the scan. In the docs, it states the first scan is supposed to do a Full sync, but no activity can be seen in the “Ingestion Jobs” tab. The “Edit Import Config” shows an error stating that the file has not been pulled with the configuration settings, when I did not even do anything to it yet. User Prop + Time Based Configuration - same issue as the first one. User Prop + Full Sync Configuration - successful creation of source and Status will show as Connected. “Ingestion Jobs” tab shows that data is indeed ingested. BUT, there are 2 problems I have observed. First, the count of rows ingested does not match the table row count in Snowflake: ingested 31,220 when the table has 38,873 in Snowflake. Some observation I had:In the “Edit Import Config” tab, time-based conf

2923
Jeremie Gluckman
Team Member
C
chris.meyerNew Member
asked in Product Experimentation
Approval System for Experiment?

Is there an approval system within Amplitude Experiment we can configure for experiment changes in a specific environment? In an ideal world I am looking for an approval system when changing targeting, rollout %, etc. where you can request reviewers for your changes and submit for approval.

2601
Jeremie Gluckman
Team Member
C
ccreaserNew Member
asked in Product Experimentation
Way to chart the time between two specific events per user?

Hello everybody!I am writing this question to see if it is feasible to perform analysis in amplitude on the period of time between two specific events performed by individual users.It would be great to be able to show the trends of this outlined time period between a user baseFor an example: say we have two events: login event sign out eventWould it be possible to chart the time period between the latest two of these respective events for each user? For example to determine how long a typical user may remain logged in before signing out?Thanks for any responses! 

6412
Jeremie Gluckman
Team Member
I
Novice
IanNovice
asked in Product Experimentation
Time-based sync for Snowflake has no jobs started

I am currently ingesting data from Snowflake, using a valid SQL query.​​​​​The sync job does not even start for the time-based sync. How long would I have to wait before the data is synced?

2999
belinda.chiu
Team Member
D
Novice
Dimple BNovice
asked in Product Experimentation
How can I view screen/ event view time in the report?

While applying the filter for events, the report gives the tap count/ view count for the screen/ event. However, instead of tap count. I want to view the time for which user has been on any particular screen.Example: Dashboard screen view is an event, and users have visited this page for 2 hr in a day and tapped for about 120 times during the day. When filter is applied for usertype = ABC & event name= Dashboard screen view, this gives tap count say 90, however, I would like to see, page view time say1.2 hr.I have tried this with ‘Segmentation’ and ‘User sessions’ both. How can I achieve the same ?

8226
D
Novice
R
richaNew Member
posted in Product Experimentation
Cannot login - how to contact support?

I can’t login - see attached screenshot.I am trying to find a way to contact support but all I see is help articles and this community. 

1852
belinda.chiu
Team Member
J
JohnNew Member
posted in Product Experimentation
Flags and variants for anonymous users

Hello I have this case where we need to rollout a feature slowly for anonymous users, but we have found out that variants doesn’t work for those users if we set them different percentages, for example I have a flag with 2 variants, A and B I set A to 90% and B to 10%, test this with an anonymous user and he get neither of those! he gets the OFF flag, anonymous users only receive a flag if I set only one variant to 100%

3504
Amplitude Admin
Community Manager
0
00rladkfk00New Member
asked in Product Experimentation
What are the exact event definition criteria for Pathfinder charts?

What are the exact event definition criteria for Pathfinder charts?(Subtitle: Pathfinder Chart vs Funnel Chart) Hello everyone, I have a question about Pathfinder charts. I tried drawing the same event using the Pathfinder chart and the funnel chart, but it occurred with completely different results on the two charts. Even if I search the guide, the exact event criteria of the Pathfinder chart are not defined. Does anyone know about the event definition criteria of the Pathfinder chart? Even if I try again and again, I get completely different results depending on the usage chart when measuring the same event. I would like to know why. Does anyone know about this case. Any help would be great for me  Thank you 😀

3191
T
Team Member

Recent Tips & Tricks

Terms & ConditionsCookie settings