Skip to content
Back to blog

Stop Exporting CSVs Every Time You Need an Answer

product-updates
workflow
integrations

The old workflow was familiar. Export a CSV. Rename it. Upload it. Realize it is already stale. Export it again.

That loop is fine once. It is a mess when it becomes the way your team answers every real question.

Live integrations change that. Anna can now start closer to the source, which means less setup, fewer one-off exports, and a cleaner path from "what is going on?" to "show me the answer."

The real problem with manual exports

CSV exports are not the enemy. They are still useful when you need a one-off file, a spreadsheet from a client, or a quick data drop from a tool you have not connected yet.

The problem starts when exports become your operating model.

Every manual export creates three kinds of drag:

  • You rebuild context instead of reusing it.
  • You lose confidence in freshness because the file is only as current as the last export.
  • You ask smaller questions because setup takes too much effort.

That is where good analysis starts to die. Not in the model. Not in the charts. In the friction before the first question.

What changed

There are now three cleaner ways into connected analysis.

1. A public integrations directory that actually helps you decide

The integrations directory is no longer just a wall of logos. You can filter by role, browse by category, and see which sources are already live versus still coming soon.

That matters because most teams do not start with "which connector is technically possible?" They start with "which tools should we connect first?"

The directory now answers that more directly:

  • public cards use compact, recognisable marks instead of stretched wordmarks
  • live integrations stay ahead of coming-soon items in each section
  • the browse experience holds up on desktop and mobile instead of collapsing into a cramped filter row

2. An in-app hub built around what to connect next

Inside the app, the integrations hub now separates four different jobs that used to blur together:

  • what is already live
  • what needs attention
  • what Anna recommends next
  • what is available but not urgent

The recommendation cards are also more honest now. They lead with what the source is for, not fluffy multi-source filler. If you are deciding whether to connect Google Ads, you should see what Google Ads gives you, not generic copy about "unlocking deeper analysis."

Recommendation UI should answer one question first: "Why this source?" Pairing advice matters after that, not before.

3. The upload modal has a shorter path out

The upload modal still lets you upload a file or start blank. That did not change.

What changed is the browse experience around it.

Instead of turning the modal into a giant scrolling directory, it now does less:

  • your live sources stay at the top
  • the best next sources are visible immediately
  • a small set of additional live sources is available inline
  • the full catalog lives behind a clear "Browse all integrations" escape hatch

That is the right split. A modal should help you decide quickly, not become a second full page stuffed into a dialog.

Better failure states matter too

Connecting a source is not only about the happy path.

Two things are much clearer now:

Plan limits

If you have reached your plan's connection cap, the app tells you before you start another OAuth flow. That sounds obvious, but silent plan-limit failures are one of the fastest ways to make a product feel brittle.

Current connection limits are:

PlanLive connectionsRows per dataset
Explore15,000
Pro525,000
Team5100,000
BusinessUnlimited500,000

Business is still 500,000 rows per dataset in the current product configuration. Not one million.

Provider setup requirements

Some integrations need more than a simple OAuth redirect. Google Ads is the obvious example. If a provider needs account-specific fields before it can connect, the frontend now surfaces that clearly instead of leaving the useful detail buried in a server log.

That is not glamorous work. It is still product work.

Why connection counts matter on pricing pages

Credits tell you how much analysis you can run.

Connection counts tell you how much of your actual stack can stay ready.

Those are different buying signals.

A founder deciding between Explore and Pro often does not ask, "How many credits do I get?" The first real question is usually, "Can I keep Shopify, Google Ads, HubSpot, Stripe, and GA4 connected at the same time?"

That is why connection allowances now show up much more directly on pricing and billing surfaces. They are not a footnote. They are part of the value of the plan.

The bigger shift

This update is not really about connectors.

It is about changing the shape of the work.

When the path into analysis gets shorter, teams ask better questions. They stop treating analysis like a scheduled project and start treating it like an available capability.

That is the goal:

  • less exporting
  • less setup repetition
  • clearer limits
  • cleaner decisions about what to connect next

Upload a file when you need to. Start blank when you want to. But when the answer already lives in a tool you use every day, you should be able to start there.

Connect the sources you already use and stop rebuilding the same dataset by hand.

Try it free

No credit card required