Angles and Offers: Why Nothing Else Matters Until $10M (And How We Build an Angle Matrix)

One brand. Two campaigns. From roughly $24k a month to north of $190k a month in a single quarter. No new product, no funding round, no clever account-structure trick.

What changed was the angles. That's it. The thing most people spend the least time on turned out to be the only thing that moved.

I've come round hard to a view that sounds almost too simple to be useful: until you're somewhere around eight figures, angles and offers are the whole game. Everything else is a rounding error. Account structure, bid strategy, the exact format, the polish on your edit. It all matters a little. None of it matters the way the angle does.

And yet I watch founders and media buyers pour their hours into all of it except the angle. So let me make the case, then show you the exact tool we use to stop guessing: an angle matrix.

Why angles sit above everything else

Here's the uncomfortable bit. Nobody has ever bought something because of your account structure.

Think about how a purchase actually happens. Someone is scrolling, half-bored, and an ad says something that lands. It names a problem they have, or a desire they hadn't quite put words to, and for a second they stop. That stop is the angle doing its job. The bid cap didn't do that. The campaign budget optimisation didn't do that. The angle did.

I believe this is why two brands can run the identical format, the identical structure, even near-identical production, and one scales to $20k a day while the other limps along. People love to blame the platform when their ads stall. In reality the account is usually fine. The ads just aren't saying anything the customer cares about.

To put a number on it: if you can find ten genuinely different angles that each work, and you feed each one a modest ~$1k/day, that's $10k/day in spend that the algorithm can actually hold. Most brands aren't capped by budget or by the platform. They're capped because they only have two or three ways of talking about the product, so there's nowhere for spend to go without the returns falling over.

That's the case for treating angles as the main event. Now the part that actually trips people up.

Concept is not angle (and confusing the two is why your ads flop)

Most people think they're testing angles when they're really testing concepts. The distinction is small to say and enormous in practice.

A concept is the format or the device. "Us versus them." "Before and after." "Three reasons why." "Founder talking to camera." These are containers. They're about what you want to make.

An angle is what's actually inside the container, aimed squarely at the customer. It's the reason the person buys. Not "us versus them" but us versus what, and on which specific thing the customer secretly cares about most.

Here's why it matters so much. You can run a "before and after" ad that completely flops, and run another "before and after" that spends $20k a day. Same concept. The difference was never the concept. It was the what. What transformation, for whom, framed how.

So when someone tells me they tested "the comparison angle" and it didn't work, my honest reaction is that they didn't test an angle at all. They tested an empty container and learned nothing. The buy happens because of the substance, not the shape. Nobody ever thought "oh good, a three-reasons-why ad, I'd better purchase." They bought because of what the three reasons were.

My rough rule: spend about 80% of your thinking on the customer and the message, and 20% on the format you'll pour it into. Most accounts have that exactly backwards, and then wonder why pretty ads aren't converting. Some of the ugliest ads I've ever seen have scaled hardest, purely because the angle underneath was right.

Right, so if angles are the game and angles aren't concepts, how do you make sure you've actually got a proper spread of them? That's where the matrix comes in.

The angle matrix: two axes, not a list

Most angle "research" produces a flat list. Twenty ideas in a doc, all roughly the same flavour, all aimed at the same person who already knows they want the thing. That's not coverage. That's one corner of the market, twenty times.

The fix is to stop writing a list and start filling a grid. Two axes.

Axis one, down the side: the angle itself. This is the core message. Each row is a genuinely distinct reason someone might buy, built from a real customer situation. Not "it's high quality" but a specific person with a specific problem or desire. A teenager with hormonal breakouts is a different angle from a forty-year-old worried about early ageing, even if it's the same serum in the bottle.

Axis two, across the top: awareness level. This is the bit almost everyone skips, and it's the half that makes the grid powerful. The same angle has to be said completely differently depending on how much the person already knows. The five stages, borrowed from old-school direct response and still the most useful lens I know:

• Unaware. They don't even know they have the problem yet.
• Problem aware. They feel the problem but don't know solutions exist.
• Solution aware. They know solutions exist but not your product.
• Product aware. They know your product, they're weighing it up.
• Most aware. They're basically ready, they just need a reason to act now.

Cross those two axes and every cell is a different ad. The "early-ageing" angle at the unaware stage is a story about a habit nobody questions. The same angle at the most-aware stage is a sharp offer with a reason to buy today. Same underlying angle, five different ads, because the person on the other end is in five different headspaces.

A flat list only ever speaks to the product-aware and most-aware crowd, because that's who's easiest to picture. The grid forces you to write for the people further up who haven't started looking yet, and that's where the volume lives.

Where to put the budget (this is the counterintuitive part)

Here's the move that surprises people: put the most budget behind the unaware end, not the bottom of the funnel.

The instinct is the opposite. The bottom of the funnel converts best, the numbers look prettiest there, so it's tempting to pile in. But that's the smallest, warmest pool of people, and it's the same crowd your retargeting is already harvesting. Scaling there just means paying to win customers you'd likely have won anyway.

The unaware and problem-aware end is where the actual ocean of new customers is. It's harder to write for, the immediate metrics look softer, and that's precisely why most brands underinvest in it and stay small. So we deliberately weight spend up the funnel: the most budget behind unaware and problem-aware creative, tapering down as the audience gets warmer, because the warm end largely takes care of itself.

I won't pretend this is comfortable in week one. Upper-funnel creative is a slower read and you have to hold your nerve while it finds its feet. But when it works, it fills the top of the funnel with people who didn't know they wanted you yesterday, and everything below it gets easier. That's the difference between an account that's stuck recycling demand and one that's genuinely creating it.

Filling the cells: research, not imagination

A grid is only as good as what you put in it, and you cannot invent this from your own head. The angle is about the customer, so the customer is where the raw material comes from.

This is the unglamorous part nobody films. Read your reviews, all of them, the five-star and the three-star. Sit in the relevant Reddit threads. Read the questions under competitor videos. Run your review dump through an AI tool and ask it to surface the personas and the wording you'd never have reached for. You're hunting for two things: the exact situations people are in when they need you, and the literal language they use to describe the problem.

Every distinct situation you find becomes a candidate row. Every awareness stage gives you a column. The cells more or less write themselves once the research is real, because you're no longer guessing what to say. You're repeating back what customers already told you, aimed at the stage they're standing in.

What this looks like put together

Say you've got a sleep supplement. A flat list would give you ten variations of "sleep better," all aimed at someone already shopping for sleep aids. Useful, but it's one corner of the room.

Build the grid instead. Rows might be: the parent of a newborn running on fumes, the shift worker whose body clock is wrecked, the anxious over-thinker who can't switch off, the frequent traveller fighting jet lag. Four genuinely different people, four angles.

Now run each across the awareness columns. The over-thinker at unaware is a relatable story about lying awake replaying a conversation, no product in sight. The same angle at most-aware is a crisp offer with a guarantee. Four angles times five stages is twenty distinct ads, each one earning its own slot, with the heaviest budget sitting behind the unaware and problem-aware cells where the new customers actually are.

That's a real testing roadmap, not a pile of near-duplicates. And the wins compound. Once a cell starts working, you iterate within it, swapping the main elements one at a time, and you've got a repeatable engine instead of a lucky hit you can't explain.

Where to from here

If you take one thing from this, let it be the reframe: stop testing concepts and start testing angles, and stop writing flat lists when a grid will show you the gaps.

Try it this week. Map your current live ads onto the two axes and see where they actually sit. I'd put money on them clustering in two or three cells near the bottom of the funnel, all talking to people who already know you. The empty cells up top are your growth, sitting there in plain sight.

This grid is genuinely the backbone of how we build creative for the brands we work with, because it turns "make more ads" into "fill these specific gaps." Plot yours, find the holes, and write into them. If you map it out and want a second pair of eyes on where the gaps are, reply and tell me what your grid looks like. I'm always happy to talk angles.