Why General Analytics Tools (Amplitude/Mixpanel) Struggle with TTV

Amplitude and Mixpanel are incredible tools. We love them. We recommend them. They are the best in the world at Event Counting and Funnel Conversion.

But if you have ever tried to build a P90 Mean Time To Value dashboard in them, you know the pain.

The "Event-Based" vs "User-Based" Problem

General analytics tools store data as a stream of Events. { timestamp: 1, event: 'click' } { timestamp: 2, event: 'view' }

They are optimized to answer: "How many times did 'view' happen?"

Time To Value is a State Change Metric. It asks: "For each unique User, calculate the time difference (Delta) between their State A (Signup) and State B (Value), and then visualize the distribution of those Deltas."

This is a computation, not a count.

The SQL Reality

To do this in raw SQL (if you export your data to Snowflake/BigQuery), you hav to:

1. Group by User.
2. Find MIN(timestamp) of Signup.
3. Find MIN(timestamp) of Value.
4. Subtract them to get TTV_Seconds.
5. Filter out NULLs (users who haven't finished).
6. Filter out Negatives (data weirdness).
7. Calculate Percentile_Cont(0.9) across the dataset.

This query is expensive, slow to run, and hard to maintain.

The "Segmentation from the Past" Problem

It gets harder. You want to know: "What is the TTV for users who were on the 'Free Plan' when they signed up?"

If you just filter by Current Plan, you miss everyone who upgraded. You need Point-in-Time properties.

• General tools usually only index user properties as they are now.
• Tivalio remembers what the property was then.

The Specialized Tool

Tivalio is the specialized screwdriver. Amplitude is the Swiss Army Knife. We pre-compute these State Changes. We model the User Lifecycle as a first-class citizen. This lets us give you P90, P50, and Spread out of the box, without writing 50 lines of SQL.