FactorPrism®
Comparison

TOP_INSIGHTS finds interesting segments. FactorPrism explains the whole move.

Snowflake's built-ins help analysts explore and retrieve. FactorPrism® locates the factors behind a metric change, quantifies each contribution, and reconciles the result to the number you need to explain.

At a Glance

Snowflake Built-ins vs. FactorPrism®

Snowflake built-ins (TOP_INSIGHTS / Contribution Explorer / Intelligence)

  • Flag or retrieve segments for an analyst
  • Flat, overlapping results that don't sum to the change
  • Awkward for metrics that cross zero (margin, net flows)
  • A SQL function or chat answer—not a finished analysis
  • You still reconcile it by hand

FactorPrism® (purpose-built native app)

  • Computes the complete answer in one pass
  • Right level including intersections; gains and declines separated
  • A breakdown that ties out to the total exactly
  • AI-written narrative, ready to hand to finance
  • Native to Snowflake—your data never leaves the account
The SQL Function

TOP_INSIGHTS Is a Decision Tree — Built for a Different Job

SNOWFLAKE.ML.TOP_INSIGHTS fits a decision tree to your data: give it dimensions, a metric, and a control/test label, and it returns the segments the tree found most associated with the difference. That's segment screening for an analyst in a worksheet. Root cause analysis — an answer that reconciles to your number, respects your hierarchy, and can be handed to finance — is not something a decision tree produces, by design:

~
Flat, overlapping segments Results don't respect your business hierarchy and don't sum to the change you're explaining. There's no bridge from baseline to actual — you reconcile by hand, and the overlaps usually mean you can't.
~
Control/test framing with a non-negative metric Natural for "this quarter vs last," awkward-to-impossible for metrics that cross zero (margin, net flows) or "explain the shape of the last 12 weeks."
~
It's a function, not a product Tabular output in a worksheet. No waterfall, no drill-down, no narrative, no way to hand it to the CFO.
The Agent

Snowflake Intelligence Retrieves Answers — "Why" Is a Computation

Snowflake Intelligence is excellent at conversational retrieval over a semantic model: "what was revenue by region last quarter?" works beautifully, and it will even tell you which region fell most. But a complete "why" is not a retrieval — it's a computation:

~
"Why" is a multi-step computation, not a retrieval A complete answer means checking every level of region × product × channel space, separating broad-based forces from localized ones, and making the pieces reconcile. Multi-step reasoning is the documented weak spot of text-to-SQL agents — you get an answer, not the complete, mutually-consistent set.
~
Semantic-model authoring Good answers require a maintained semantic layer — real data-team work before business users get value.
~
Unpredictable cost Consumption-priced chat means the bill scales with curiosity.
The Finished Analysis

The Problem FactorPrism® Was Built to Solve

A reconciled root cause analysis has never existed as a product anyone could install — analysts have always assembled it by hand. FactorPrism® is a Native App that computes the complete answer in one pass: no SQL, no semantic model, no prompt engineering.

+
Reconciles exactly A waterfall bridge from baseline to actual where ranked factors sum to precisely the change. Every dollar of movement accounted for — the standard a finance audience already holds you to.
+
Finds where causes live Hierarchy-aware: a broad-based force acting across the whole business is reported once, at the top — not smeared across every leaf segment. A localized problem is pinned to the exact segment where it acts, with when it started ("step change beginning March 12").
+
Handles real finance metrics Movement across all periods (not just endpoints) and sign-crossing metrics like margin and net adds.
+
Knows when not to speak A significance gate validated against ground truth — benchmark published keeps the list short and trustworthy — the documented failure mode of this category is alert fatigue, and we built against it.
+
Stays in your account Same as the built-ins: it runs on your warehouse, under your governance. Your data never leaves Snowflake.
Side by Side

The Ingredients vs the Finished Dish

Your business isn't a set of disjoint segments. What happens to it is a set of overlapping causes — each acting at its own place, some everywhere at once, some inside one region's slice of one product line. Segments are where causes show up; they're not what happened. That's the difference in one sentence: tools built to slice show you slices, and FactorPrism® returns the causes, located where they act, reconciled to your number.

The Built-ins

  • TOP_INSIGHTS: SQL function, segment table output
  • Intelligence: chat agent over a semantic model
  • Flat segments — no hierarchy, results don't sum to the change
  • Non-negative metrics only (function); answers vary (agent)
  • Audience: SQL-writing analysts
  • Building blocks — you assemble the story by hand

FactorPrism®

  • Native App with a business-user interface
  • Reconciled waterfall + ranked factors + narrative
  • Hierarchy-aware: broad-based vs localized, split explicitly
  • Movement shapes, sign-crossing metrics, dated onsets
  • Setup: a read-only grant on your data
  • One reconciled answer, ready for the CFO
FAQ

Honest Answers

Can't we get this from TOP_INSIGHTS?

No — and not because it's a lesser tool, but because it's a different kind of tool. A decision tree flags segments associated with a difference; it doesn't locate causes in your hierarchy, separate broad-based forces from localized ones, or produce a bridge that sums to the change. If you've run it on a real variance question, you've seen the gap: a flat list of overlapping segments that doesn't tie out to the number you're explaining. Closing that gap is not a bigger decision tree — it's a different computation, and it's the one FactorPrism® performs.

Is this competing with Snowflake?

No — FactorPrism® exists because of Snowflake's platform. It's a capability you add to your account, reviewed through the Marketplace security process, billed through your Snowflake relationship.

What does FactorPrism® need access to?

A read-only grant on the tables or views you choose to analyze — nothing more. Optionally CORTEX_USER (AI-written explanations) and CREATE DATABASE (saving sessions in your own account).

Can I see it without connecting data?

Yes — the in-app demo replays a complete analysis on a realistic retail dataset, instantly.

See the difference on your own numbers

Watch the product demo, request a guided evaluation, or install from the Snowflake Marketplace to find the factors behind your KPI change.

Get it on Snowflake Marketplace