The Kalicube Opportunity Matrix: three columns, and you do them in order
Bruce Henderson gave the industry a grid. His Growth-Share Matrix at BCG plotted what a company already owned against how fast the market was growing, and it produced stars, cash cows, dogs and question marks. The brutal question underneath it was a good one: what do you already have, and is it worth feeding?
The industry heard “matrix” and drew boxes. McKinsey made nine of them, and consultants have been adding cells ever since.
I’ve kept Henderson’s question and thrown away his grid. The Kalicube® Opportunity Matrix is three columns, and you do them in order.
Does Exist. Should Exist. Could Exist.
You don’t plot a page in a box, because a plot lets you start anywhere, and starting anywhere is the leak. The machines don’t plot your brand either. They read what’s already there, then they notice what’s missing, and only then are they in any position to reward something new. That’s a sequence, so the work is a sequence.
Does Exist is where the value already is, and it’s the column everybody skips
What you already have. Audit it, clean it, optimise it, and remove or redirect whatever drags the brand down.
The risk here is zero. The investment is already sunk, the proof already exists, and you’re recovering it rather than betting again. That’s Return On Past Investment, and it’s the cheapest return in the business.
Most of the value sits in this column, and it rarely looks like a content problem. It’s the Entity Home that contradicts the LinkedIn company page. It’s the article from 2019 that still ranks and still describes the offer you stopped selling. It’s the interview where somebody introduced you with a job title you’ve since outgrown. It’s the comparison table you published as an image, which is perfectly readable to a human and close to nothing to the machine that has to parse it. It’s the case study sitting in the CRM that nobody ever turned into a page.
That last one isn’t a hole in the content calendar, it’s a harvest, and it’s the Codified gate doing exactly what it’s for. Harvest it out of operations, codify it to one source of truth, distribute it across first party, second party and third party. The outcome was already true. Nobody had written it down where a machine could read it.
Before I write anything new for Kalicube, I ask one question: does a live page already answer this, badly?
Should Exist is the gap your competitors are already standing in
This is what competitive analysis and the machines between them show to be missing. A question your ideal customer is already asking that you’ve never answered. A claim an assistive engine can’t ground because nothing outside your own site repeats it. A conclusion a cheap model won’t infer, because you never stated it.
A contradiction between two pages you already own isn’t a hole, it’s mess, and mess belongs to the first column. Nobody should have to audit the same contradiction twice.
The risk is low, because you didn’t invent the gap. The data revealed it, and the Funnel Query Pathway tells you which node to fill. Do not invent the gap. An invented gap is a Could Exist wearing a Should Exist costume, and it’ll cost you a quarter.
This is ordinary Return on Investment: present spend, present return, a known timeline. You build the missing page, then you get the missing corroboration, then you earn the missing proof. Claim it on the Entity Home, corroborate it somewhere you control but don’t own, and prove it somewhere you can’t author at all.
Could Exist is where nobody in your category has thought to be
I’ve been sloppy about this column in the past, and I want to correct it. Could Exist isn’t “what you invent.” That definition is too small, and it sends people straight to the blank page when the blank page is usually the wrong move.
Could Exist is where your digital footprint could expand beyond where it currently sits. Should Exist is the ground my competitors are already standing on, so I should be there too. Could Exist is me looking at somewhere I could be present that isn’t obvious to anybody, including them.
It has three faces, and only one of them is content.
Where you could be. Take what you already have and put it somewhere your category hasn’t reached. A platform, a surface, a format of publication, an audience that’s genuinely yours and that nobody has bothered to serve. Nothing gets written. Something gets placed.
What could exist. The original thought, and this is the part I’ve had wrong. Original doesn’t mean built from nothing. You take the data you already had, or the perspective you already held, and you make something from it that didn’t exist before: a framework, a novel angle, a conclusion nobody else has articulated. The raw material comes out of the first column. The object is new. Repurposing what you already know is Could Exist work, not housekeeping, because the output is a thing that wasn’t there.
Who could carry you. The partner, the collaborator, the complementary business whose audience overlaps yours and whose voice isn’t yours. This is the only one you can’t author, which is exactly why it belongs here. No competitive analysis in the world returns a partnership that doesn’t exist yet.
Where, what and who. Place, thing, voice. Their formal names are the channel, the creative and the partnership, so that’s three things under three names rather than six types.
All three are built on something that already exists, which is why this column comes last rather than first. It’s also why it works on the machines. Novelty and agreement get weighed together, and the industry calls the first half information gain. A conclusion pitched against everything your market holds reads as an outlier and gets discounted for it. The same conclusion resting on ground the market already agrees on reads as a contribution. You need something solid to stand on, and Does Exist is where you go and find it.
The three publication tiers run inside this column as well, and they’re what turn a Could into something else.
I publish the new thing on my own site first, because until I do it only could exist. The moment it’s live, it does exist, and that isn’t a figure of speech. It has changed column. Then I want a second party to corroborate it somewhere I control but don’t own, so the thing I made is confirmed by somebody who isn’t me. Then I want a third party, who owes me nothing, to establish that the thing existing is meaningful rather than merely true.
Exists, confirmed, meaningful. First party, second party, third party.
Notice what that does to the columns. A Could Exist I published on Monday is a Does Exist by Tuesday, and if nothing outside my own estate has confirmed it, it’s a missing proof, which is Should Exist work. So these aren’t three buckets you sort your material into once. They’re three states the same item moves through, and it comes round again. That’s a further reason a grid can’t hold this: a grid has no way to say that a cell you filled last quarter is a different cell now.
The risk is higher in all three, because demand is unproven, and that’s the point of doing them third. A Could Exist piece sitting on a dirty Does Exist portfolio trains the wrong story, because the machines read the mess first and the brilliance second. When you place dated public proof early and the world catches up to it later, that’s Return on Latent Proof, and it only pays if the rest of the record is clean when the world arrives.
The types sit inside the columns, and they never cross
There’s no three by three here, and I want to be exact about why, because I’ve seen this sold as a grid and it isn’t one.
Each column holds its own named types, nested inside it, and not one of them is an axis.
Does Exist holds the underperforming asset, which exists but is weak, so you fix the content. The unformatted asset, which exists but isn’t ready for an algorithm to read, so you add structure. The underamplified asset, which exists and is fine and nobody ever distributed it. And the digital inconsistency, where two surfaces you already own contradict each other and an engine has to pick a winner. Every one of them already exists, which is precisely what puts them in this column.
Should Exist holds the coverage gap, a topic you simply missed. The displacement, where a competitor occupies a slot that should be yours and you go and reclaim it. And the missing proof, where your claim is sound and nothing outside your own estate has ever confirmed it. Every one of them is a hole.
Could Exist holds the three faces above, under those same formal names: channel, creative and partnership.
A grid would let you enter at any cell. A tree makes you walk the trunk first, and that difference is the entire method.
You run the same three columns on one page or on a whole territory
Asset and territory aren’t two frameworks, they’re two zoom levels on the same three columns. You can run Does, Should, Could on a single article, and you can run it on an entire topic, a channel, or your presence across the language models. Same order every time.
It isn’t a sixteenth gate of The Kalicube Framework either. The fifteen gates describe what happens to your material once it’s out there. This describes the order you work on the material before you let it go.
The build order is the whole method, and the timings are honest
Find the asset or the outcome that’s already true. Harvest it, codify it to one source of truth, distribute it. Build the missing node the data points at. Then, and only then, go somewhere nobody has been.
Skipping the first column doesn’t only leave a mess for the machines to read before they reach your good work. It starves the third column of the material that column is made from.
For a personal brand it’s about a year: three months on Does, three on Should, six on Could. For an SMB or an enterprise you’re looking at two and a half years and often more: six months, then twelve, then twelve to twenty-four. Smaller companies are frequently slower than big ones, and it’s resources rather than velocity that slows them.
The mistake is always the same, and it’s a billing problem
Agencies skip Does Exist because an audit is harder to sell than a creation. SEOs start at Could because a new piece feels like progress, or at Should with a gap they’ve talked themselves into. Henderson at least made you look at the portfolio before you spent anything, and we’ve kept that. We’ve just made it a sequence instead of a plot.
Title the Matrix. Speak the three columns. Do them in that order, or the third one leaks.
The Kalicube Opportunity Matrix was coined by Jason Barnard in 2025.