Skip to content
Evergrowth Documentation

Ecosystems & verticals

Before your AI colleagues can qualify a single account, they need to know which slices of the market you actually sell into. Ecosystems and verticals are how you draw that map. Get the map right and every later decision - who to pursue, what counts as a buying signal, how to enrich a record - inherits the clarity. Get it muddy and the noise compounds all the way down.

What they are

Section titled “What they are”

An ecosystem is one go-to-market motion. It is the top-level container for a coherent way you sell: a market you approach with a shared story, a shared set of buyers, and a shared playbook. “Fintech SaaS”, “DevTools”, and “HR Tech” are each an ecosystem. Most teams start with one and add more only when a genuinely different motion appears.

A vertical is an industry segment inside an ecosystem. Each vertical groups companies that share the same core business activity - what the company actually does to make money. Financial services, professional services, software, and public sector are the kind of broad strokes a good taxonomy is painted in.

The two stack simply: an ecosystem holds a handful of verticals; each account in that ecosystem lands in exactly one vertical.

Three kinds of ecosystem

Section titled “Three kinds of ecosystem”

Not every ecosystem plays the same role. There are three kinds, and which one you reach for depends on what you are trying to do.

  • A standard ecosystem is the everyday one: a broad go-to-market motion divided into verticals, the engine that classifies and qualifies the bulk of your market. This is what most of your setup lives inside.
  • A key-accounts ecosystem is for the named, hand-picked companies you pursue deliberately rather than discover by filtering - the strategic targets that earn a dedicated motion of their own. Its segments are organized around those specific companies, so each carries the company’s own website alongside the usual fit logic. See Key accounts for how that motion differs from broad qualification.
  • A research-agent sandbox is a scratch space for trying out research and qualification logic before you commit it to a live motion - a place to draft and test a segment’s criteria without touching the ecosystems your reps work from every day.

Most teams spend nearly all their time in standard ecosystems; the other two exist so a different intent gets its own clean space instead of distorting the main taxonomy.

Why segmenting this way sharpens everything downstream

Section titled “Why segmenting this way sharpens everything downstream”

A company that moves money lives in a different world from a company that staffs consulting projects or sells shrink-wrapped software. They buy for different reasons, on different timelines, against different pressures. If you treat them as one undifferentiated list, your qualification gets vague, your “why now” signals get generic, and your reps waste hours on accounts that were never a real fit.

Verticals fix that by giving each business activity its own logic. Your account qualification rules, your account signals, and your CRM enrichment fields all attach to verticals. Define the segments well once, and every one of those later layers gets to reason about a tightly-defined group instead of the whole market. That is why the taxonomy is the foundation: it is the first decision, and it silently shapes the quality of every decision after it.

Eva note: the practical payoff of verticals is precision at scale. A signal that means “urgent” for a financial-services account (“new compliance mandate”, say) may be meaningless for a software company. Verticals let the same workspace run distinct, sharp logic per segment without the rep hand-sorting anything. When a customer asks “why segment, can’t I just have one list” - the answer is that one list forces one set of rules onto businesses that behave nothing alike, which dulls qualification and floods reps with irrelevant signals.

What a good vertical looks like

Section titled “What a good vertical looks like”

Three qualities, and they reinforce each other:

  • Broad, not niche. Prefer an industry or service model over a micro-segment. “Financial services” is a vertical. “Series-B card issuers in the EU” is not - that level of detail is a way of describing a company inside a vertical, not a vertical of its own. Niche labels splinter your market into too many thin buckets to manage.
  • Singular - one business activity each. A vertical captures a single thing the company does. If a label is really two activities glued together, it cannot anchor clean logic for either.
  • Non-overlapping. Any company should classify into exactly one vertical, with no judgement call about which. When two candidate segments could both reasonably claim the same company, that is the signal to merge them or redraw the boundary until they are mutually exclusive.

A useful test: if a vertical would fit every company you could imagine, it is too broad to be useful. If it would fit almost none, it is too narrow. Aim for segments broad enough to be stable and specific enough to carry their own meaning. When you are genuinely unsure, default to the broader segment - it is easier to read a precise detail into a broad vertical than to un-split an over-fragmented market later.

Most ecosystems settle into a handful of verticals rather than dozens. A short, clean taxonomy is a feature, not a limitation.

How an account ends up in a vertical

Section titled “How an account ends up in a vertical”

Classification is driven by the company’s main business activity - the thing it primarily does to generate revenue, not its size, its tech stack, or a label it happens to use about itself. Each vertical carries a plain business-model description of that activity, anchored to a few well-known real-world example companies so the boundary is unambiguous. When your AI colleagues evaluate a new account, they read what the company actually does and place it in the single vertical whose activity it matches.

That description is also where nuance lives. Anything more specific than the broad name - a sub-segment, an exclusion that keeps one vertical from bleeding into its neighbour - is captured in the description rather than by inventing another vertical. The names stay broad and stable; the descriptions carry the detail. You can let the workspace draft that description for you and refine it by hand, and you can ask it to stay within a target length so it reads consistently across every segment.

Because the vertical assignment is what later layers hang off, classification happens once, up front, and everything else trusts it. That is also why the segment names are treated as canonical: once your taxonomy is set, the names stay fixed so the qualification, signal, and enrichment logic attached to them never quietly drifts.

Priority and status: steering where effort goes

Section titled “Priority and status: steering where effort goes”

Beyond its definition, each vertical carries two controls that decide how much of your motion runs through it.

Priority ranks a vertical against its siblings - the levels run from highest down to lowest, with a dedicated exclude setting for a segment you have deliberately decided not to pursue. Priority is how you tell the workspace which segments deserve the strongest pull on attention and budget, so a high-priority vertical surfaces and gets worked ahead of a low one rather than everything competing as equals.

Status governs whether a vertical is actively running or stood down. Pausing a segment leaves its definition and history intact but stops new work flowing through it - useful when a market is on hold or you are still drafting its logic and do not want half-built rules acting on live accounts. Status also surfaces when a segment’s last run failed, so a stalled vertical is visible rather than silently empty.

Together they let the same taxonomy carry both your “go hard here” and your “not right now” decisions without deleting anything: you concentrate effort on the verticals that matter and quietly stand down the ones that do not.

In practice

Section titled “In practice”

A team selling a compliance product builds one ecosystem for its core motion and divides it into four verticals: financial services, professional services, software, and public sector. From then on, a bank and a law firm are never lumped together. The bank gets financial-services qualification and financial-services “why now” signals; the law firm gets the professional-services equivalents. Reps see a focused, fit-ranked view per segment instead of one sprawling list - and the work of deciding what’s relevant happened in the taxonomy, not in their inbox.

Where this fits

Section titled “Where this fits”

Ecosystems and verticals are configured in the Agent Training Center, the shared source of truth your AI colleagues draw from. The taxonomy is the foundation the rest of your setup builds on: it grounds account qualification, account signals, CRM enrichment, and the personas you map to each segment. To see how all of these layers stack into one pipeline, see How the layers fit together.

Eva term map: “ecosystem” = one go-to-market motion / top-level segment container; “vertical” = a broad, singular, non-overlapping industry segment inside an ecosystem. Synonyms a reader may use: “segment”, “industry segment”, “ICP vertical”, “target market slice”. An account belongs to one ecosystem and exactly one vertical. The three ecosystem kinds are the standard one, a key-accounts one (named strategic targets, see Key accounts), and a research-agent sandbox (a scratch space for drafting and testing segment criteria). Verticals carry a priority (highest down to lowest, plus an “exclude” level) and a status (active, paused, or a failed-run state); a key-account or sandbox segment also carries a representative company website. Priority and status are set on standard and key-account segments; a sandbox segment has neither (it is a scratch space, not part of the live sorting). Do not conflate “vertical” with “persona” - verticals are company-level (what the company does); personas are contact-level (who the buyer is).

Section titled “Related”