A practical guide

9 min read

The scouting process for corporate innovation teams

Most scouting does not fail at the search. It fails at the brief that was too broad, the longlist nobody finished reading, and the shortlist that could not be explained six weeks later. This is the method that avoids those, phase by phase.

What is innovation scouting?

Innovation scouting is the structured search an organisation runs to find external partners that can solve a specific business challenge. It moves from a clearly stated challenge, through a broad longlist of company profiles, to a small shortlist evaluated against agreed criteria. Done well, it is repeatable and its decisions can be explained later.

Note what the definition does not say. It does not say the partner has to be a startup. The right answer to a challenge is sometimes a young company, and just as often an established supplier, a research institute, or a business in an adjacent industry that solved the same problem years ago. Deciding the type of partner before you understand the challenge is itself a scouting mistake.

The definition sounds tidy. The practice rarely is, and the gap between the two is where most of the value is won or lost. The rest of this guide is about closing that gap.

On this page

The failure modes

Why scouting stalls

Scouting projects rarely collapse in a dramatic way. They stall quietly, in five familiar places. Recognising them is most of the work, because each one has a fix that costs almost nothing if you apply it early.

The brief that matches everything
When the challenge is written as a theme rather than a need, the search returns the entire adjacent industry. There is no basis to exclude anyone, so the longlist balloons and the scout ends up ranking on instinct.
The longlist nobody reads
A longlist is only useful up to the point a human can process it. Past a few dozen entries in a spreadsheet, rows stop being opened. The list exists, the coverage is claimed, and most of it is never actually seen.
Evaluation by inconsistent standard
Scoring spread across a shared spreadsheet drifts. Each rater weights different things, uses the scale differently, and fills the gaps with judgement nobody wrote down. The totals look objective and hide the disagreement.
The shortlist you cannot defend
Six weeks after the search, a sponsor asks why a company was rejected. If the reasoning lived in someone's head or a deleted tab, the honest answer is that no one remembers. A shortlist you cannot reconstruct is not a decision, it is a guess with good formatting.
Knowledge that leaves with the scout
When the search, the criteria, and the ratings live on one person's laptop, they walk out the door when that person changes role. The next project starts from zero, and the same field gets scouted again as if for the first time.

The method

Thirteen steps, five phases

A complete scouting project runs through thirteen steps, from the first statement of the business challenge to reusing what you learned on the next one. Listed flat, thirteen steps read like a compliance checklist. In practice they cluster into five phases, and the phase is the useful unit: each one has its own job, its own definition of good, and its own way of going wrong.

The phases hold whether you run them in a spreadsheet or in software.

01

Frame the challenge

  • 01Business challenge
  • 02Search field identification
  • 03Acceptance criteria

Everything downstream inherits the quality of the framing. Before a single search runs you need to know what the business actually needs solved, which angles are worth pursuing, and what would make a company worth talking to.

What good looks like. A good challenge names the outcome and the constraint. “We need partners who can help us cut battery pack weight without raising unit cost” gives a scout something to push against. From there the field breaks into separable search directions, each with its own vocabulary, and each direction gets acceptance criteria that a company can genuinely pass or fail.

The mistake to avoid. Writing the challenge so broadly that everything matches. “We are looking for sustainability innovation” is not a challenge, it is a category. Anything can be argued into it, which means nothing can be argued out. The same failure repeats whenever criteria are left until after the longlist is already open.

In Coopsaas

Describe the challenge in plain language. Coopsaas drafts the search directions and acceptance criteria from it, so your team edits rather than starts from a blank page.

02

Search the field

  • 04Guided search
  • 05Company discovery

With the framing agreed, the search becomes mechanical rather than exploratory. Each direction is translated into structured parameters, and you work through matched company profiles to build a longlist with honest coverage.

What good looks like. Good scouting keeps the longlist scannable. You are not reading every profile in depth, you are making fast include or exclude calls against criteria you already agreed, then going deep only on the candidates that survive.

The mistake to avoid. Collapsing everything into one keyword and treating the first page of results as the market. You end up scouting the terms you already knew and missing the adjacent field where the best answer usually lives. The other version of this is a longlist of three hundred rows that nobody opens past row forty. Coverage you cannot process is not coverage.

In Coopsaas

Search directions become structured search parameters against an integrated database of company profiles, so the longlist arrives ready to triage instead of pasted in from ten browser tabs.

03

Evaluate together

  • 06AI-assisted evaluation
  • 07Team collaboration
  • 08Shortlisting

Scoring is where a shortlist earns its authority. Each candidate is rated against the agreed criteria, ideally by more than one person, so the result reflects the team rather than whoever happened to run the search.

What good looks like. Everyone scores against the same criteria, disagreements stay visible instead of being averaged away in silence, and every rating carries a short reason. Six weeks later you can still say why Company A beat Company B.

The mistake to avoid. Evaluation done in a spreadsheet where each person quietly applies a different standard. One rater scores on traction, another on technology, a third on gut feel. The numbers look comparable and mean nothing.

In Coopsaas

AI-assisted evaluation enriches and scores profiles against your criteria with the reasoning always visible, and the team rates and comments in one shared workspace.

04

Decide and prove

  • 09Partner selection
  • 10Proof of concept
  • 11Innovation partnership

A shortlist is not the deliverable, it is the setup for a decision. This is the handover to a proof of concept or a first conversation, and the moment the case for each finalist has to stand on its own.

What good looks like. Good handover carries the reasoning with the name. The sponsor sees not just who made the shortlist but why, and what each candidate would need to prove in a pilot. That is what turns a list into a defensible recommendation, and a successful pilot into a partnership worth scaling.

The mistake to avoid. Treating selection as the finish line. A shortlist with no clear next step, no owner, and no pilot criteria tends to sit until the momentum, and often the budget, is gone.

In Coopsaas

The reasoning behind every score travels with the shortlist, so the case for each finalist is ready to hand to a sponsor.

05

Keep what you learned

  • 12Monitoring
  • 13Repeat

The last phase is the one most teams skip, and it is the one that makes the next project cheaper. Markets move, and the work you just finished is an asset only if you keep it somewhere the team can reach.

What good looks like. The search directions, the criteria, and the rejection reasons stay in a shared record. When the same field comes up again, the next project opens the last one instead of starting from zero. Companies you passed on stay tracked, because “not yet” often becomes “now” a year later.

The mistake to avoid. Letting the project end with the slide deck. When the search, the criteria, and the ratings live on one person's laptop, they walk out the door when that person changes role, and the same field gets scouted again as if for the first time.

In Coopsaas

Search directions and acceptance criteria are saved and reusable, and monitoring keeps selected companies and markets in view after the project closes.

The part everyone gets wrong

How to write acceptance criteria that discriminate

A criterion that everyone passes is not a criterion. It is a description. “Uses modern technology” and “is innovative” feel like standards, but no company you are looking at will fail them, so they do no sorting. A useful criterion is one where a real candidate could plausibly fall on either side of the line.

The test is simple. For each criterion, name a company you would happily talk to that would still fail it. If you cannot, the criterion is not doing any work. “Has a live deployment with a reference customer in our industry” passes that test. “Has strong technology” does not.

Agree the criteria before you open the longlist. This is the single most important sequencing rule in scouting. Once you have seen the candidates and quietly picked a favourite, the criteria start to drift toward whatever that favourite happens to be good at. Set them first and they measure the field. Set them after and they only confirm a decision you already made.

Keep scoring consistent by making reasoning visible. A team will never apply a scale identically, and trying to force that is a waste of everyone's time. What you can do is require a short reason beside every score, and put every rater's scores in the same view. When two people rate the same company three points apart, that gap becomes a conversation instead of an average that hides the disagreement.

A quick sense check

  • Each criterion has a candidate you like that would fail it.
  • The whole team saw and agreed the criteria before the search.
  • Every score carries a one-line reason.
  • Scores from different raters sit side by side, not merged.
  • The weighting is written down, not held in someone's head.

A checklist you can apply

How to tell a good shortlist from a long one

Length is not quality. A good shortlist is short because things were deliberately taken off it, and every name that stayed can be defended. Run yours against these.

  • You can say in one sentence why each company is on the list.
  • You can say why the obvious candidate you left off did not make it.
  • Every company on it clears the criteria you agreed before the search, not criteria invented afterwards.
  • More than one person looked at it, and their reasoning is on record.
  • A sponsor who was not in the room could read it and understand the shortlist without you narrating it.
  • There is a clear next step for each name: a conversation, a pilot, or a specific thing to prove.

A fair comparison

How teams run scouting today, and when each approach fits

There is no single right answer here. The best choice depends on how often you scout, how much of the reasoning you need to keep, and whether you are building an in-house capability or buying a one-time result. Most teams end up combining several of these.

  • Spreadsheets and email

    The search, the longlist, the scoring, and the decisions are tracked by hand across spreadsheets, slides, and email threads.

    Where it is strong
    No cost, no procurement, and complete control over your own method. Nothing new to learn if the team already works this way.
    Where it strains
    There is no workflow logic, so every project rebuilds the structure from scratch. Coverage is capped by how much one person can read, and the reasoning behind decisions scatters across files.
    The right choice when
    A one-off search narrow enough that a single scout can hold the whole picture in their head.

  • A company database subscription

    A paid subscription providing searchable company profiles, used alongside whatever tracking the team already has.

    Where it is strong
    Reach. You surface companies you would never have found through your own network or a search engine.
    Where it strains
    A database gives you a list, not a process. There is no evaluation layer, no pipeline, and no collaboration, so the hard part of the work still happens somewhere else.
    The right choice when
    Your process is already solid and the gap you need to close is purely coverage.

  • A CRM adapted for scouting

    An existing sales CRM repurposed with custom fields and pipeline stages to track scouting projects.

    Where it is strong
    Already procured, already approved by IT, and often already familiar to part of the team.
    Where it strains
    The data model is built for managing customer relationships, not for evaluating partners against criteria. The workarounds hold for a while and then break as the team grows.
    The right choice when
    Rarely a deliberate choice. It is usually a stopgap picked for procurement reasons rather than fit.

  • General-purpose AI chat

    An assistant used to research companies, summarise profiles, and draft comparisons on demand.

    Where it is strong
    Fast at summarising, and genuinely useful for exploring an unfamiliar field or drafting a first set of search directions.
    Where it strains
    It cannot hold a methodology, run an evaluation across a team, or keep a record of decisions. Answers vary between sessions and nothing accumulates.
    The right choice when
    Early exploration, before the process starts. Treat the output as a starting point rather than a shortlist.

  • An external consultancy

    A specialist firm runs the scouting for you and delivers a report with a recommended shortlist.

    Where it is strong
    Outside networks and dedicated capacity for a burst of work, with a polished deliverable at the end.
    Where it strains
    You inherit their method rather than build your own, and the knowledge leaves when the engagement ends. Cost and turnaround make it hard to run often or iterate quickly.
    The right choice when
    A high-stakes search where an outside network matters more than building an in-house capability.

  • A structured scouting workflow

    Software that combines the company database with the process itself: challenge, search directions, acceptance criteria, shared evaluation, and a logged record of every decision.

    Where it is strong
    The method is repeatable and the reasoning is retained. Repetitive steps are AI-assisted, and the whole team works from one shared view instead of separate files.
    Where it strains
    There is a workflow to adopt and criteria to agree on. The value compounds across projects, so a single throwaway search will not show it off.
    The right choice when
    A team that scouts more than once and needs decisions it can defend and knowledge it can keep.

Frequently asked

Questions about the scouting process

Is scouting only about finding startups?

No, and framing it that way narrows the search before it starts. The right partner for a business challenge might well be a startup, but it might equally be an established supplier, a research institute, a university spin-off, or a company in an adjacent industry that has already solved the same problem. Good scouting starts from the challenge and stays open about where the answer comes from.

How long does a scouting project usually take?

It depends on how tightly the challenge is scoped and how many search directions you run. A focused search with clear criteria can reach a shortlist in a couple of weeks, while a broad, exploratory field takes longer because the pruning and the reading both take longer. The framing decisions in the first phase drive the timeline far more than the size of the database.

How many companies should a longlist contain?

Enough to trust your coverage, few enough that a person will actually read all of them. A longlist that no one can process is not more thorough, it is less. Tighter criteria and better search directions matter more than raw volume.

When should acceptance criteria be set?

Before anyone sees the longlist. Criteria defined after you have a favourite tend to bend to justify it, and the evaluation stops measuring the market and starts measuring your first impression. Agreeing them up front is what keeps a rejection defensible later.

How do you keep scoring consistent across a team?

Share one set of criteria, use the same scale, and require a short reason with every score. Consistency comes less from a clever formula than from making each rater's reasoning visible, so that disagreement is discussed rather than silently averaged into a number.

What is the difference between a longlist and a shortlist?

A longlist is about coverage: the full set of company profiles that plausibly fit, gathered so you can be confident you did not miss the obvious candidate. A shortlist is about commitment: the small group you have evaluated against your criteria and are prepared to take into a conversation or a pilot.

How do you avoid losing scouting knowledge when someone leaves?

Keep the search, the criteria, the ratings, and the reasoning in a shared record rather than on individual laptops. If the next project can open the last one and see what was searched, what was rejected, and why, the knowledge stays with the team instead of the person.

Run this process without the spreadsheets.

Coopsaas turns the thirteen steps above into one scouting workflow, with an integrated company database and a record of every decision your team makes.

See how the product worksBook a demo