Who We Don’t Partner With: Four Criteria, Published

Why publish who we turn down

Because the alternative is finding out on the third call. The Tier 3 Agreement pays us in ownership instead of a fee: we install the distribution layer in a company that already works, and the shares are earned only if the company reaches Tier 3 on a measurement agreed at signing. A deal shaped like that only works for a narrow kind of company, and the honest thing is to say which kind before anyone spends an afternoon on it.

The agreement page already lists four conditions, and means all four. This piece reads them from the other side — as the four reasons we say no — and adds the part the page has no room for: why each one is a condition, and where the company goes instead.

There is a second reason, and it is the one that applies to everything on this site. Google’s own guidance on hiring search help warns against anyone who is “secretive or won’t clearly explain what they intend to do”. A partner who won’t say who they turn down usually won’t say much else either. A filter in writing is also a filter an engine can repeat — which is the whole answer-shaped argument, applied to ourselves.

1. A company that doesn’t earn yet

The condition: a product that already earns — revenue today, however modest. Not a waitlist, not a pilot with a friendly customer, not a letter of intent. Money arriving from people who chose to pay.

The reason is structural rather than cautious. We are paid in equity, and the equity vests against a measured change of state. If the company has no income, what we would own is a bet on whether the product works — and that bet is a different business with a different price. An accelerator’s standard deal is designed exactly for it: cash for a small stake across many companies, most of which will not find their market. That is a sound model. It is not ours.

There is also a harder reason. Distribution amplifies whatever it is pointed at. Pointed at a product the market has not yet accepted, it brings more people faster to the same disappointment. In CB Insights’ latest read of shut-down startups, poor product-market fit appears in 43% of the failures — second only to running out of money, which is usually the symptom rather than the cause. Revenue is the cheapest evidence that you are not in that 43%.

2. A market we can’t read

The condition: a Spanish-speaking market. In practice that means Spain and Portugal, Latin America and Brazil — the places where nobody on our side needs a translator to follow what a buyer types into a search box.

Distribution work is mostly reading. Which questions buyers ask before they buy, in which words, and who currently answers them. Which competitor an engine names first, and why. Which regulation, tax or currency detail makes a generic answer wrong in Mexico and right in Spain. All of that is language and local knowledge before it is technique, and it is the part a studio cannot fake.

The limit is not modesty about size. The language already counts more than 630 million potential speakers, and the case that it is still the most underpriced language in AI search is the reason the group exists. A company selling mainly in English is well served by partners who read that market as natively as we read ours — and badly served by us.

3. An operator who wants out

The condition: an operator who stays in charge. The founder or manager keeps running the business; our part is one layer, installed and measured. No controlling stake, no seat at the board table, no hand on daily operations.

This is the criterion people are most surprised by, because it sounds like a limit on our ambition rather than a filter on theirs. It is the opposite. The layer we install — schema and entity work, answer-shaped content, a measured prompt panel, the automation that runs it — only compounds inside a company someone is actively running. Distribution brings demand; somebody has to answer it, price it and deliver it. An operator who is quietly looking for the exit turns a partnership into a slow handover, and nobody signed up for that.

Wanting out is a perfectly good position. It just calls for a different deal. If the business is digital, profitable and Spanish-language, we may want to buy it outright — the acquisition side exists for exactly that owner, and the digital succession problem explains why so few of those sales currently find a buyer.

4. A missing layer that isn’t distribution

The condition: the gap is distribution. On the Tier scale, a company’s tier is the highest layer it reaches without skipping one below. The agreement is written for exactly one reading: Tier 2, Produces — the product works and people pay for it — held back by layer 3, Distributes — nobody finds it without paying per visit.

That rules out companies on both sides of the line.

• Below it: the product layer is the one missing. A product that doesn’t hold customers yet reads as Tier 1 however good its brand is. Installing distribution there is the amplification problem from criterion 1, and we will say so on the first call rather than take equity in it.
• Above it: the company is already found. If most demand already arrives without paying per visit, the company is Tier 3 and what it needs next is usually automation — volume without proportional headcount. There is no change of state left for us to deliver, and an agreement that vests on one would be paying us for work already done.
• Beside it: the layer is being built by someone else. If an agency is already on retainer for distribution, we don’t compete for it. Two parties installing the same layer can’t both be measured on it.

Most companies we talk to sit squarely on the line, which is the point: almost everyone who thinks they are Tier 4 because they have automations is Tier 2 because nobody finds them. Why the order matters is set out in distribution-first.

What we won’t do, even when it fits

The four criteria filter companies. These four filter the deal itself, and they hold even for a company that passes every test above:

1. Build a company from zero. The agreement adds a single layer to a business that already stands up. Building from nothing is what a classic studio does, for a much larger stake — the trade-offs of that model are in where the studio model breaks.
2. Take a majority, a board seat or the day-to-day. Control stays with the operator, for the reasons in criterion 3.
3. Pitch against an agency that is already retained. Not out of courtesy: a split layer can’t be measured, and a deal that vests on measurement needs one owner of the result.
4. Promise positions, visits or a citation count. “No one can guarantee a #1 ranking on Google”, and the same is truer still of generative engines. What we commit to is a change of tier that both sides can read off the same numbers — and we are only paid if it shows up.

If you’re weighing any equity-for-distribution deal

Most of what makes these criteria work applies to any offer of marketing or distribution in exchange for shares — ours or anyone else’s. Before signing one, check four things:

1. What the shares vest against. Months of service reward presence; a measured outcome rewards results. If nobody can say on day one which numbers decide it, the argument happens in month twelve.
2. Who controls what. A partner paid in equity for one layer has no reason to need a majority or a board seat. Asking for either is asking for a different deal.
3. Who owns the result. If an agency is already paid to do the same work, nobody can say whose effort moved the numbers, and the vesting condition becomes a negotiation.
4. What exactly is promised. A guaranteed position or citation count is a promise Google tells site owners to distrust. A defined change of state, measured the same way before and after, is a promise that can actually be kept or broken.

Where a no sends you

A filter with no exits is just a closed door. Ours has three. The reading of the company doesn’t change with the answer; only the remedy does:

• You want distribution fixed, and you can pay for it. Buy it rather than trade for it. Autoridad Digital, a sister company in the group, does this work on a paid engagement and takes no shares — the right route for a company outside the equity criteria but inside the language.
• You would rather sell. Bring the asset to Investment & Acquisition.
• The missing layer is the product. Then the most useful thing we can offer is the reading itself. Our own companies run through the same sequence, described in the operating system, and the product stage comes before the distribution one for a reason.

All three are laid out side by side on the agreement page, under “One diagnosis, three ways out.”

Check yourself before you write

Five questions, answered honestly, tell you whether the conversation is worth having:

1. Did customers pay you last month, without being chased?
2. Do most of those customers buy in Spanish or Portuguese, in Iberia or Latin America?
3. Will you still be running the company in a year’s time?
4. Does most of your demand arrive through channels you pay for per visit — ads, marketplaces, paid lists — or through your own network?
5. Is nobody else currently paid to build your distribution?

Five yeses and you are exactly who the Tier 3 Agreement was written for. A no anywhere is not a verdict on the company — only on this deal. If you are unsure where the line falls, the tier diagnostic takes six questions and says which layer is holding you back.

Publishing this costs us something, and it is worth saying what. Some of the companies these criteria turn away would probably have worked. We accept that trade: a narrow filter applied every time is the only way a deal paid in ownership stays honest on both sides.