Getting Data for Analytics

What do you want to know?

Collaborate with: Product Management

• What do we maybe want to count?
• What stories will we want to tell?
• What product decisions are we making in the future?
• Put this in writing with a URL on it

What do we have in a database?

Collaborate with: Engineering

• Is there already a data model, defined schemas of what they plan to save in a database?
• Can it answer your questions?
• Can it change slightly to answer more of your questions?
• Get this synced to the data warehouse and start using it.
• Work together to document the columns

Will server-side events help?

We take the lead.

Collaborate with: Engineering

• Write schemas, try to cover all the questions
• Use the review process to discuss if all the fields are available or if there's more stuff you should get

Do you have some Click Tracking you can use?

Collaborate with: Analytics, Data Engineering

If you have any data that gives referrer, page pairs you can use this to infer a lot about clicks rather than instrumenting an event on each link.

• Create a table (view, whatever) that is just the page/referrer info you need for this project
• Be especially careful with your regexes to identify those pages. (Counting things is difficult)
• Get reviews from both Analytics and Data Eng if you can

Can you answer all your questions?

Collaborate with: Product Management

• What questions still aren't covered?
• Can they be reframed in a way that is answerable from what we do have?

Browser events, if you have to.

Please keep in mind that there will be some event loss due to adblockers and network latency. Because the loss is due to user choice, it is not random.

Collaborate with: Engineering

• Write schemas
• Use the review process to discuss if all the fields are available or if there's more stuff you should get