How To Save Your Scaleup

A little over a week ago, I released Scouut’s Mid-2026 Startup & Scaleup Market Memo. If you haven’t read it, you should. If you have, and you work in or around the scaleup world, you probably came away worried.

I don’t want to sugarcoat this. You should be worried.

Most of the scaleups I can see are on course for something close to a catastrophe. Many still have growing revenue, happy customers, strong brands and metrics that would have made investors very happy just a couple of years ago. Internally, they look healthy. But new startups coming from behind are reaching the same milestones with far fewer people, much less capital, and often in less time too. Taking a boat from England to America was once the best, most effective way to make that trip - until it wasn’t.

There are silver linings though. Scaleups have customers, distribution, capital, company knowledge and more - very real advantages. And most importantly, I believe there is just enough time for any given scaleup to be saved, providing they are willing to do what’s required.

What’s interesting is that the required action, however different and extreme it feels, is just the result of applying core scaleup building principles that have been true for years. Some could even argue nothing’s changed at all, depending how far upstream you’d like to look.

The Culture Deck

Netflix’s Culture Deck is one of the greatest documents ever written on how to build a successful scaleup. It was created and developed by Patty McCord, along with Reed Hastings, the co-founder of Netflix, who joined at roughly 30 people as Chief Talent Officer, stayed 14 years, and left with the company at a little over 2,000. She is regarded as one of the greatest team and company building experts in the world.

You might wonder why a document first published in 2009 has any relevance in a world where everything has changed in the previous 6 months alone. For a lot of people, the Culture Deck won’t be new at all - it’s famous now, and gets discussed in boardrooms all over the world.

But it isn’t a slideshow full of answers, or an instruction booklet to follow. It’s a model Netflix applies constantly to navigate, deal with, and overcome whatever they’re facing in the most effective way possible. Think of it as a compass, or the map that you’d use when you don’t know the exact directions. Whatever the catalyst for change - a financial crash, a global pandemic, or AI - the model guides the way.

The Answers Are Changing

The Culture Deck holds a number of values. Netflix believe, correctly, that a company’s values are not what it says when asked, or what’s written on a wall, but what its decisions reveal.

The deck is why some of these values are so well known, but the core ideas in it far pre-date Netflix as a company. So while I want to give it credit and tip my hat, we’ll leave it there and focus on the ideas themselves, not the deck or the company.

Think of these values as the most effective way to build a world beating company through scaleup phase, rather than the unique values of any one company.

There are actually 7-ish core values, but they can basically be summed up with the following:

• Hire the best people possible, part ways with everyone else (talent density)
• Pay them well
• Reward impact, not effort
• Provide enough context and ensure company wide alignment

What matters most is that these values do not change when the external climate does. They are fixed. So when something like a pandemic, or AI, comes to town, decisions get made the same way, through the same model.

That does not mean making those decisions have the same answers. If the conditions have changed, and you run them through the same model and values, you are almost guaranteed to get different answers out the other side. This is ultimately what we’ll be proving today.

As we go through each value and how AI as a catalyst is redefining what the correct answers are, it’ll become clear how deeply connected and intertwined they all are.

Talent Density

Hiring the best people and parting ways with everyone else leads to what is known as talent density. The term gets used to describe lots of things, which is why I prefer my longer, albeit less catchy explanation.

Talent density is a measure of how much exceptional talent you have per seat. It’s one of the most important things any scaleup must optimise for, and a lack of it immediately causes a ripple effect of other problems all through the business.

Common scaleup complaints, such as red tape, bloated process, layers of approval, an appearance of busy work, can nearly all be traced to a dip in talent density. Some of this is necessary as a company scales, but when density is high, the need for it comes far later, at much larger scale, than when density is low or average.

Most innovative and venture backed companies believe they have high talent density. Few truly do. And while many could improve the caliber of people they bring in, most struggle with density not because of who they hire, but who they keep.

Companies often see top performers who may be 2, 3 or even 5x more capable than their lowest performers, but choose to retain both, which is a mistake. Talent density requires you to hire high performers, and to keep the gap between your highest and lowest performers small.

The only way to do that as a company scales quickly is by removing lower performers. I say lower, not low, because against the rest of the market these people may still be well above average. But if they don’t clear the bar for what’s acceptable at your company, even if they once did, they no longer meet the requirement to remain.

AI has moved that bar substantially. As I said in the memo, top performers are breaking away from the rest, so the advantage of having them in your company is as big as it’s ever been - and the disadvantage of not having them is just as big. That alone means your team needs to be reassessed against the new bar. Many teams that legitimately had high density in January are without it today, even if the team hasn’t changed at all.

AI has also altered the individual talent level of each person you employ. However you assessed talent before - a marketing role, an engineering role, or anything else - you’d have assessed for the skills that the role required at the time. Almost all roles now require a level of AI fluency, and to be excellent at most, AI expertise is the requirement.

The chance that everyone employed by a scaleup in January 2026 now clears that new bar is 0%. There is no “some chance”, no “maybe, if” - the chance is nil.

So if people have not left the team in the past ~7 months, by their own choice or the company’s, you do not have the talent density needed to succeed in the new world, and core principle and value of maintaining a high density team is no longer being met.

People will fight this idea and argue with it. They’ll say “your employees have families, they may enjoy working for you, some of them may need to work for you to provide.”

But letting someone go because they sit below your bar is not the same as deciding they’re incapable of finding work. Your bar should be one of the highest in the world. Not meeting it doesn’t mean a person’s life falls apart because of a decision you made. And if you genuinely worry they can’t get another job quickly, or that nobody else will hire them, that’s all the more reason to ask why they’re still on your team.

Yes, it’s unfortunate. No, it isn’t evil, and you are not ruining a life. Part with them the right way - Netflix is open about its strong severance packages. You don’t hate them, you don’t want them to fail. Their job, at the level they can do it, just no longer exists in your company.

It’s impossible to keep a garden from overgrowing and running wild without routine pruning of the plants that are out of place. The conditions have changed, and some of your plants can no longer thrive. That’s it.

Pay Top of Market

Paying top of market shouldn’t be confused with paying competitively, which just means paying at a level where you’ll have lots of competition. Paying top of market means the offers you make to new hires, and the salaries of the people you choose to keep, are in line with what the top end of the market would pay them.

These numbers will, obviously, be above market average. And “market” does not mean the market you’re in - if you’re a Series C scaleup, you don’t benchmark against other Series C scaleups. Market means the realistic market for the talent you want to hire. One of my clients is in Australia, but the realistic market for the type of staff they hire is Anthropic and OpenAI in San Francisco - so that’s what they benchmark against, not local Aussie startups. Market is the competing market for potential new hires, and for the high performers you already have.

You need to pay top of market because the best people are unlikely to join, and likely to leave, roles where they aren’t paid it. The alternative is hiring a caliber of talent with no top of market offers pursuing them, which moves you further from talent density, not closer.

Like talent density, paying top of market isn’t a principle or value that has changed. But the figures required to keep meeting it have.

As the bar has risen and the pool of acceptable talent has shrunk, the relative demand for that talent has grown. Simple supply and demand tells you these people will be more expensive. But supply and demand isn’t the only variable. Potential ROI plays just as big a role in what someone, or in this case “market”, will pay. If I expect something to return $10, I might pay $8. If I expect something else to return $10,000, I’ll obviously pay far more than $8, even if it’s readily available and supply is no issue.

So, fewer people worth employing than 6 months ago, more competition for them, and a far higher expected return from having them.

Numbers vary from ecosystem to ecosystem, but we’ve seen top performers double their salary overnight within the past few months. Multiple payrises of 20-30% in February, and then again in July, would not be unheard of right now. In most cases I’d argue that’s sane spending. I’ve written more on how companies should think about salary spend here, but the maths isn’t hard to run in your own business, and it will tell you that paying top of market to maintain talent density is far cheaper and friendlier to burn than your other options. By comparison, the numbers on those options now look so bad it’s hard to see how companies doing anything else survive.

And as companies that don’t adapt begin to drop, ‘market’ goes through its own form of density increase, so the numbers behind paying top of market will keep evolving.

Like talent density, if you haven’t given your best performers a payrise in the last ~7 months, the chance they are simultaneously at the bar required for talent density and paid top of market is also 0%. If you can’t justify a bump for a top performer, it’s usually a sign they need to be let go, and that your bar simply isn’t high enough for this new world. A real top performer has dramatically increased their impact and ROI over the past 7 months.

Reward Impact, Not Effort

Impact and effort get treated as if they’re the same thing, particularly by less experienced managers, or by anyone who has never worked with someone truly exceptional.

Effort is how much work someone does, and how hard they appear to be doing it. Impact is what actually changes in the business because of that work. The two absolutely have some correlation - high impact people often put in a fair amount of effort too - but plenty of people put in enormous effort for no meaningful impact, and some create enormous impact with very little effort at all.

There are exceptions, but in a team environment the person with the most impact and the person with the highest amount of effort are rarely the same person.

AI has dramatically increased the impact one person can have, so the definition of ‘high impact’ has changed. It varies from function to function, and large scale impact also relies on other parts of the business being in order (as you’ll see shortly), but for some people AI has multiplied their impact more than tenfold. These people have to be recognised and rewarded, because if you don’t, rest assured somebody else will.

Their visible effort is unlikely to be rising in the same fashion, though - it may even be falling. I know people who’ve legitimately 8-10x’d their impact and now work 9am-4pm because their brain is fried if they go any harder. That should be rewarded. Working until 8pm adds no commercial benefit if it comes at a cost to the impact they provide.

Meanwhile, others use AI to produce tonnes of work with heaps of effort. But effort without impact should not be rewarded, and people who consistently fail to have impact should be moved on, or you risk talent density issues. Praising and rewarding them also actively drives your genuine high impact performers away.

If you don’t have a method to objectively measure impact, not effort, at the individual level - and it’ll vary by role and business - it’s impossible to actually reward impact. And if you’re promoting, praising and encouraging high effort where there’s no impact, as many managers and cultures do, you’re doing the very opposite of what builds a high impact company.

Impact Needs Context and Alignment

It’s effectively impossible to have a large amount of impact without the context to create it. Many scaleups struggled to provide enough context even pre-AI, which is why so many are plagued with meetings to repeatedly provide it, and why so many managers spend their time answering questions that only exist because context is missing.

AI now expects each individual to have the impact of a small team, which means they need the context of a small team. Where one person might sit in a single meeting a day where they only need to pay attention for 15 minutes, a team might sit across 3-4 meetings, needing to absorb all of the context available.

You can argue - and I would - that most of those meetings were never necessary, just cover for low talent density. But either way, in today’s world a meeting is rarely the answer. If context isn’t documented and available on demand, you can’t expect people to reach the level of impact the company now needs to win. If they’re constantly stopping to ask, or waiting on others who hold the other pieces, you can’t get the most out of your high impact people. The companies furthest along are engineering context (literally called context engineering) into the systems they run on, so agents have permanent access to it and their human orchestrators can operate at full impact.

Be wary, though. Even when context is provided, low talent density teams will ignore or misunderstand it - they can’t be trusted to turn context into useful impact, which is exactly why you need density first. If you can’t trust your people to act on context alone, you almost certainly have a talent issue. Or an alignment one.

A well paid, high density team with all the context they need can still have very little impact if they aren’t aligned on where that impact should be directed. Alignment is closely linked to context - you can’t have it without context - but where context is the information and resource needed to do a task in the moment, alignment is about the goal and the outcome. It’s what lets people decide for themselves whether the work they could spend time doing is worthwhile, without having to ask or be told, and it stops them spending time on things that only look worthwhile at face value.

At Nike, the marketing team is aligned on building a brand and a feeling, not selling a specific item, so nobody wastes time on a poster for a new water bottle even though, logically, it might help them sell water bottles. That same alignment is what lets you hand your high impact people problems to solve and goals to hit, without needing to tell them exactly how to get there.

Without alignment, great people will do high quality but ultimately useless work. You’ll also find a company that’s plagued with politics, fallouts and more that you’d really prefer to do without.

The values in this whole piece are themselves something a company has to be aligned on, which is exactly why Netflix published the Culture Deck publicly, and why people confirm they’re aligned to its principles before they join. Your managers and your people will need to understand and share these values too, or you’ll have little chance of ever successfully achieving them.

Yes, it is extreme

I said in the memo that the required change would feel extreme. This piece is titled How to Save Your Scaleup for a reason - something that’s fine doesn’t need saving, but most scaleups today are certainly don’t look fine when you can see the competition that’s not all that far behind them.

Living by these values, and executing within them, looks very different today than it did a little over 6 months ago - and the correct answers, whatever they are for your company, will probably be extreme.

• Wanting talent density used to mean parting with maybe 10% of the team. Today it may be 40%.
• Paying top of market used to mean frequent but small increases. For some roles it now means paying 100% more tomorrow.
• Rewarding impact might mean promoting someone who works 30 hours a week whilst parting ways with someone who does 70.
• Teams that used to have meetings constantly might barely meet once a week, and roles that existed to move context around may no longer be needed at all.

While the specific answers absolutely do depend on each specific company, none of this is pure speculation. There are scaleups already in the middle of saving themselves, or that likely already have, and this is directionally what it looks like.

As I said at the beginning of this piece, I think there is just about enough time - but that window is closing very quickly, and I’m not confident that many will do what’s needed to squeeze through until it’s too late…