Wave 0 — Ecosystem and actor mapping

Before researching the domains from first principles, map the people and organisations already engaged in the same quest: who they are, what they do, what they publish, and who they refer to.

Why this comes first

Four distinct yields, in increasing order of long-term value:

1. A shortcut to knowledge. Others have already aggregated much of what the domain map asks for. Reading a research lab’s ten years of output is faster than rediscovering it.
2. Source discovery at compounding rate. “Who do they refer to” is the engine. Each actor cites others; following citations reaches the field’s real centre far faster than searching does. This is snowball sampling, and it is the most efficient discovery method available for a field you do not yet know.
3. A permanent monitoring base. The output is not a one-off report but a watchlist — a set of feeds, journals, newsletters and channels we follow from now on. This is the difference between learning a field once and staying current in it.
4. A landscape, including its white space. Who is doing what, where, and — the part that matters commercially — what nobody is doing. Our founding observation is that very few entrepreneurs push these solutions. This wave tests that claim honestly instead of assuming it.

Two further benefits that fall out for free: a pipeline of potential partners, suppliers, collaborators and clients; and a read on what the field considers credible, which tells us what we will have to prove to be taken seriously.

Why it must not come alone

Two risks, both real:

• Inherited blind spots. Map a field by following its own actors and you adopt its consensus, including what it has collectively failed to question. Actor mapping tells you what the field believes; it does not tell you what is true. It informs the domain briefs; it does not replace them.
• Evidence tier. Most actors are marketing themselves. Companies and influencers are Tier D/E by default — excellent for discovering what exists, unreliable for what works. Universities and research labs are the exception and often route to Tier B.

And one sequencing point that matters more than anything else in this document: a twelve-month baseline you did not start is the one thing you cannot catch up on later. Instrumenting Site Zero has a long lead time and near-zero intellectual overhead — it should start now, in parallel, and log quietly while this desk research runs. Do not let a rich, interesting ecosystem-mapping exercise consume the season in which you could have been collecting your first year of data.

Actor taxonomy

The three categories proposed, plus five that will surface anyway and are worth naming in advance:

T7 deserves special attention early — a single conference programme or fair exhibitor list is the cheapest actor-discovery instrument that exists.

Geographic rings

Each ring is searched for a different reason, not merely for more of the same.

R5 is a genuine structural advantage and should be worked hard rather than treated as a footnote: native-language depth, an existing network, and a city you can be in for a day without it being a trip. R1 is where actionability is highest and published information is thinnest — expect R1 to be fieldwork (conversations, site visits, fairs) rather than desk research, and budget time accordingly.

Method: seed, profile, snowball, saturate

  ┌─ 1 SEED ──────────────────────────────────────────────────────┐
  │  30–50 candidate actors from obvious entry points:            │
  │  conference programmes, association member lists, agency       │
  │  grant records, "who we work with" pages, existing contacts   │
  └──────────────────┬────────────────────────────────────────────┘
                     ▼
  ┌─ 2 TRIAGE ────────────────────────────────────────────────────┐
  │  score each; decide: PROFILE / MONITOR / LOG / DISCARD        │
  └──────────────────┬────────────────────────────────────────────┘
                     ▼
  ┌─ 3 PROFILE ───────────────────────────────────────────────────┐
  │  actor profile for each PROFILE-grade actor:                  │
  │  what they do, what they publish, what they claim, evidence   │
  └──────────────────┬────────────────────────────────────────────┘
                     ▼
  ┌─ 4 HARVEST REFERENCES ────────────────────────────────────────┐
  │  extract every actor they cite, cite them, partner with,      │
  │  are funded by, or appear alongside → back to step 2          │
  └──────────────────┬────────────────────────────────────────────┘
                     ▼            ▲                                 
                     │            └── repeat until saturation ──────┘
                     ▼
  ┌─ 5 SYNTHESISE ────────────────────────────────────────────────┐
  │  clusters, hubs, white space, watchlist, source registry      │
  └───────────────────────────────────────────────────────────────┘

Step 1 — Seed

Cheap, high-yield entry points, roughly in order of efficiency:

• Event programmes and exhibitor lists — one document, fifty actors, already filtered.
• Association and cooperative member directories.
• Public grant and funding records — who received money for what, with amounts. Often the single most honest map of a field.
• University department pages and lab publication lists.
• “Partners”, “clients” and “as featured in” pages on company sites.
• Existing personal network — including the finca community around you; ask who they have hired, read, or been impressed by.

Aim for 30–50 seeds spread deliberately across all eight categories and all six rings, not 50 startups.

Step 2 — Triage

Score 1–5 on four factors, then decide:

Recording discard reasons is not bureaucracy — it is what stops the snowball from looping.

Step 3 — Profile

One actor profile per PROFILE-grade actor. The sections that do the real work:

• What they publish, and where — every channel, because each is a candidate feed.
• What they claim, and what evidence they offer — tier it. An actor making strong claims with no method is itself a finding.
• What they do not cover — the white space is assembled from these.
• Who they reference — the fuel for step 4.

Step 4 — Harvest references

For every profiled actor, extract every other actor they cite, partner with, are funded by, or appear alongside — and record the direction of each reference. This produces a network, not just a list, and the network is more informative than any individual entry:

• An actor referenced independently by many others is a hub. Profile hubs deeply; they are the field’s load-bearing structure.
• An actor nobody references is either marketing, or genuinely novel. Both are worth knowing, and they look identical until you check.
• Clusters of mutual reference reveal the field’s schools of thought — and, usually, which ones do not talk to each other. The gaps between clusters are often where the interesting work is.

Keep the reference edges as data (from → to, type of reference). Even a plain list of edges supports counting who is cited most, which is most of the value.

Step 5 — Saturate, then synthesise

Stop rule: when two consecutive snowball rounds produce no new PROFILE-grade actors. Not when the list feels long enough. Expect saturation in three to four rounds per ring; R1 will saturate fastest and R5 slowest.

Time-box: six to eight weeks of part-time work, reviewed at the halfway point. If it is going to run longer, that is a decision to take explicitly, against the cost of not doing domain research.

Synthesis deliverables are listed below.

Deliverables

The last row is the one that justifies the wave. If this exercise produces a beautiful map of the ecosystem and nothing flows into the domain briefs, it was networking, not research.

How this changes the research sequence

The domain map’s prioritisation proposed starting with instrumentation, then pools, then water. This wave inserts ahead of it — with one non-negotiable concurrency:

  NOW ──┬── Wave 0: ecosystem & actor mapping        (desk + fieldwork, 6–8 weeks)
        │
        └── Site Zero instrumentation begins          (runs in parallel, starts now,
             — the baseline clock starts               logs quietly for 12 months)

  THEN ──── Wave 1: P1 measurement · A1 pool · W1–W4 water · context · E4 energy
             now informed by everything Wave 0 surfaced

Wave 0 makes Wave 1 much cheaper. It does not replace it.