1. Overview

Source: content/4segments/manual/01-overview.md

1.1 What 4Segments is for

4Segments answers one question over and over: "Which of my records match this combination of conditions, and how many are there?"

You answer it by dragging things onto a canvas and clicking the part you want. The number updates as you click. When the number looks right, you send those records somewhere — a list in your marketing automation platform, a CSV on an SFTP server, a spreadsheet on your desktop.

The point of doing this visually rather than by writing a query is that you see the overlaps. A query tells you the answer you asked for. A Venn diagram also tells you about the 746,946 records in the part you didn't ask about — which is often the more interesting number.

Typical uses:

  • Building an audience for a campaign or nurture track
  • Producing a call list for a sales rep, scoped to a territory
  • Sizing a market before committing budget to it
  • Finding out what actually happened to a segment since last month
  • Understanding where your data is thin, dirty, or surprising

1.2 The six words you need

Almost everything in this manual is built out of six ideas. Learn these and the rest follows.

Datasource — a database 4Segments can query. You may have several: your 4Comply system, a Snowflake warehouse, a BigQuery dataset, a MySQL database. Each one is configured once by an administrator.

Criterion (plural: criteria) — one column in one table of one datasource, given a friendly name. SF Industry is a criterion. SF Job Title is a criterion. Think of a criterion as a question you can ask about a record.

Criteria value — one specific answer to that question, defined by a condition. Technology is a criteria value of SF Industry. 25 – 49 is a criteria value of SF Lead Score, defined as LEAD_SCORE between 25 and 49. Each value carries a record count.

Canvas — a working surface with one diagram on it. A canvas has a type (Venn, Matrix, Waterfall, Funnel or Map), and you can have up to six open at once as tabs.

Segment — the records you have currently selected on a canvas. Click one lens of a Venn diagram and you have a segment of 405,264 records. Click a second region and the segment grows. A segment is what gets exported.

Saved segment — a segment you gave a name to. It becomes reusable: drop it on another canvas as a circle, use it to filter a map, or make it the root of a waterfall. It is a definition, not a frozen list — its size is recalculated, so a saved segment named "Hot MQLs" always means today's hot MQLs.

1.3 The shape of a session

A normal working session looks like this:

  1. Open a new canvas and pick a visualization.
  2. Drag two or three criteria values onto it from the sidebar.
  3. Look at the numbers. Adjust.
  4. Click the region you want. Check the count.
  5. Open the Records list to sanity-check that the rows look like what you expect.
  6. Export, or save the segment for later, or drag the selection to a fresh canvas and keep narrowing.

Step 6's last option is the one people underuse. A Venn diagram holds three circles; that is a hard mathematical limit, not a product limitation. But a selection dragged onto a new canvas arrives as a single circle, and you can then add two more criteria to it. Repeat as many times as you need. There is no ceiling on how many conditions a segment can ultimately represent.

1.4 One datasource per canvas

Everything on a single canvas must come from the same datasource and the same table. 4Segments enforces this: once you drop the first item, tables from other datasources grey out in the sidebar, and an attempt to mix them raises:

Everything on a canvas must come from the same datasource.

The reason is that a segment resolves to a single query against a single connection. 4Segments does not join across databases. If you need to combine two systems, that combination has to exist in the data before 4Segments sees it.

1.5 What 4Segments does not do

Being clear about this up front saves time:

  • It does not modify your data. 4Segments issues read-only queries. Nothing you do on a canvas writes back to your database.
  • It does not clean your data. The Normalization feature makes two values count as one for segmentation purposes. It does not correct the underlying rows.
  • It does not manage users. Accounts, passwords and roles live in your identity provider. See Chapter 12.
  • It does not join tables. One canvas, one table.
  • It does not give different users different rows. Everyone in a workspace sees the same data. Access is one of two levels: standard user, or administrator.

Next: 2. Getting started →