Data Planning and Governance

Data (Beta) provides data planning and tracking that helps companies run on clean and consistent analytics by eliminating human error during the instrumentation process.

259 Topics
631 Replies

MayaTeam Member

published in Data Planning and Governance

Can Amplitude track marketing attribution?

You can use the Marketing Analytics Browser SDK, which extends the Browser SDK with automatic web attribution and page view tracking. Our developed docs include the configuration related to web attribution and page view tracking here. For other functionality, check the Browser SDK. Comment below with any questions!

Andrii DatskivNew Member

asked in Data Planning and Governance

Change Amplitude "User ID" for old users

We plan to change the logic of the Amplitude user property "User ID" setup. Previously, we used to pass the value from our backend system there, and now we will start passing the firebase user id. We are planning to rewrite the value of this user property to firebase.user_id for old users (i.e. for all those who will have the "User ID" value before the logic is changed). Can our analytics "break" from changing the values for old users?

lizzyroNew Member

posted in Data Planning and Governance

Default properties not showing up

Default properties listed in the help docs below such as Device Family, Country, etc. are not displaying in my org’s instance of Amplitude. How do I turn these on?https://help.amplitude.com/hc/en-us/articles/115002380567-User-properties-and-event-properties

trevin-bcPeer Moderator

asked in Data Planning and Governance

When shoudl you use flat list of event properties vs object arrays?

Amplitude supports object arrays as event properties as described herehttps://help.amplitude.com/hc/en-us/articles/9623000954907-Cart-analysis-Use-object-arrays-to-drive-behavioral-insights#terms-and-concepts that combined with property splitting seems powerful. However I’m unclear when it’s best to use this vs a flat list of properties. Is it strictly when you need to represent multiple objects and thus would have repeating of properties forcing you to come up with creative naming to avoid collisions?in our case, say you only even have one subscription plan in a purchase. Each plan say has a name, term length and price. Also given customers can change plans, we’d want to know the previous plan they are coming from. is it just better to model event properties like:Event: Plan Purchasedplan_name: silver plan_price: 19.99 plan_months: 1 plan_name_previous: goldor is it better to do:Event: Plan Purchasedplan: { [ name: “silver”, price: 19.99, months: 1 ] } plan_name_previous: goldI ass

JeffNew Member

posted in Data Planning and Governance

My event source is always unknown.

I used @amplitude/marketing-analytics-browser@0.4.0 for tracking my events.When I track my event, the Source of event is always “unknown”.I set Browser SDK Source and I followed the initialization guide faithfully.I set the API_KEY with the key inside the Browser SDK Source page. I don't know why Source is in unknown state. Is there something wrong with my settings?

Richs86New Member

asked in Data Planning and Governance

Accidentally deleted a project, support bounces back

Who do I need to reach out to in order to get a project restored? Have tried submitting a ticket but it keeps bouncing

trevin-bcPeer Moderator

asked in Data Planning and Governance

What are best practices for property naming?

Wondering what is the best practice for property names and if there are advantages to normalizing them across events.For example, imagine I have Event Foo and Event Bar. Foo is an event involving a product and say Bar is an event involving a category. Which is the best course of action in data modeling in Amplitude:Option 1:Event Foo Property: product_name Event Bar Property: category_name Option 2: Event Foo Property: name Event Bar Property: name Basically what I’m asking is if there is any benefit to normalizing property names for charting or any other insights? Or is it a better practice to be as specific as possible? The upside of this latter approach is that when we look at our property list in our data management, it’s really clear what a given property is for without needing to dig into the description.The downside is that there are now N number of properties where as I could instead have a single “name” property for multiple events.

joshNovice

asked in Data Planning and Governance

Why can't I see the details for invalid events?

This question was asked before, but not properly answered. I set up my project’s tracking plan and implemented the event tracking in our code base. In theory, this should ensure that there are no invalid events (i.e. events being sent that don’t match the tracking plan). However, this does not appear to be the case. Many of my events are being marked as invalid and I have no idea why. Is there a way to look specifically at events that have been marked invalid so I can debug the issue causing them to be marked as such and remedy it? I’m thinking of something like the “User Look-Up” tab with the event stream where I can see events and their associated users as they come in. I can then click on a user and view the specifics of each of their events. What I want is to have this exact view of individual events but for events marked invalid, so I can see what is wrong and fix it.

MayaTeam Member

published in Data Planning and Governance

Do you need to send a time-stamp with Amplitude events?

For a client side event via the Amplitude SDKs you don’t need to worry about a time-stamp. For the server side events you need to have a timestamp attached Amplitude events if the event was performed back in time. If you want to keep the time as loaded, whenever you send that event to Amplitude the timestamp will be added. You can learn more about sending the timestamp if you want to backfill events here.

MayaTeam Member

published in Data Planning and Governance

How do user properties apply to an event?

When a user fires an event that is captured by Amplitude, it includes the current values for each user property, but these values can change over time. Learn more about how Amplitude updates user properties in our Help Center and keep us posted in the comments if you have any questions. This helpful video is also a great place to start.

MayaTeam Member

published in Data Planning and Governance

How do you keep your product data in the EU?

We can host the data on the EU data center or the US data center:If you’re interested in having your data in the EU data center - you can sign up here - https://amplitude.com/get-started-eu If you’re interested in having your data in the US data center - you can sign up here - https://amplitude.com/get-startedAdditional ResourcesAmplitude’s Stance on Security & Privacy Data Processing Addendum for Terms of Service Privacy Notice

MayaTeam Member

published in Data Planning and Governance

How do you set GDPR-compliant user IDs?

The General Data Protection Regulation (GDPR) has always been top of mind for Amplitude, and our platform does not require companies to provide their customer’s personal information (PII.) As a general rule, We recommend NOT sending any PII/ non-compliant with privacy regulation data to Amplitude. Here’s more info on identifying your users correctly and our approach to identity reconciliation.We welcome the opportunity for these conversations as we all work together to address ever-evolving GDPR requirements.Additional ResourcesAmplitude’s Commitment to Privacy and the GDPR Amplitude’s Stance on Security & Privacy

alex.hoTeam Member

posted in Data Planning and Governance

When should you use “Array” vs. “Enum” as a property type in your Tracking Plan?

Within your Amplitude Tracking Plan, you can set to ensure that your property values adhere to your planned data expectations. One way is to set specific property types that dictate the format of the property values sent to AmplitudeUsers often ask when is the best time to use “Array” vs. “Enum” as a property type?It’s important to note that that arrays and enums are not interchangeable. An array accepts a list of items while an enum accepts a single item from a predefined list.From a data health perspective, I would always lean towards using enums where possible in your tracking. The options are finite and therefore if an incorrect value is sent in, Amplitude can surface the error and flag it in the Tracking Plan UI.Lastly, you can always update the enum at anytime to include new values if you are expecting the list of values to change.For more information on how to use the Tracking Plan, check out our article here.

msnyderNew Member

asked in Data Planning and Governance

Invalid Event Details

We have an event that is working the majority of the time, but there are the occasional “invalid” events coming in as well. Is there any way for us to view the details of these invalid events? I’d like to be able to see what was sent for these specific events so we can troubleshoot the core issue.

Amplitude AdminCommunity Manager

published in Data Planning and Governance

Will subsequent calls override the property if you set a new user ID?

Generally, subsequent calls override the property if you set a new value. With user IDs specifically, the platform operates a little differently. Once a user ID is set the same one should be set for subsequent events. You should not be setting individual user IDs for users because, ultimately, you will be creating new unique users. If you decide to populate the user IDs yourself, it should stick, but this is not a best practice. For a user property (that refers to specific attributes), once you set those properties in Amplitude, they will persist for the next set of events until you set a new value. And so those new values will override that last set value from the last call.An Amplitude differentiator is that we keep the historical values of every user property, so that when you query in the past, we are able to return the value the user property had at that point in time.For helpful definitions of the following terms, refer to our glossary here. User property Event property User

Kay LeeNew Member

asked in Data Planning and Governance

Which is better event name?

According to amplitude academy, fundamentals of data taxonomy design course, they recommend using past tense verb. (see the bold & underlined text)Here are some best practices for creating a taxonomy that makes your data easy to learn:Maintain a consistent style (i.e. case and punctuation) for all of your events and properties. For example, you may use Title Case for all events and snake_case for all properties. Regardless of what you choose, the most important thing is to stay consistent. For easy differentiation, you should try using one style for your events and a different one for your properties. For events, maintain a consistent syntax that includes a noun and a verb. For example, all of your events may follow noun + verb or verb + noun order. This will help you and other data users better map what event is triggered when an action is done in the product. Name properties such that you can intuit their data type. While this may not always be possible, consider naming your prop

tiboNew Member

asked in Data Planning and Governance

Event warning "Your org has exceeded the 10M event limit" despite plan update

Hi, I recently got an email for a event volume limit hit with the object "Your org has exceeded the 10M event limit" despite the recent plan update from Amplitude stating we now have “unlimited events”. Am I doing something wrong? Did I miss something? This scares me a lot as we don’t want to have our account blocked. Thanks