The Three Games of Growth Engineering

Bottom Line:

• Growth engineering runs three distinct games: getting seen, converting attention to action, and building systems that compound. Each has different metrics, time horizons, and ceilings.
• Most teams over-invest in the attention game because it's easy to measure. The persuasion and systems games have higher ROI but diffuse credit.
• The right priority stack runs roughly in reverse order: systems first, then persuasion, then attention. Most teams do the opposite.

Growth teams fail in a specific way. They work hard, ship fast, and hit a ceiling. Then they pour more budget into the same channels and wonder why the numbers won't move.

The reason is almost always the same. They're playing one game when they should be playing three. Understanding what growth engineering actually is helps clarify why.

Game 1: Attention

Attention is the game of getting seen. Paid channels, organic reach, targeting, creative. You know this game well - it's the one with the clearest metrics. Impressions, clicks, CTR, ROAS. You can trace a line from spend to signup.

That legibility is exactly what makes it dangerous. Companies spend 6x more on acquisition (Game 1) than on conversion optimization (Game 2), despite conversion work having higher ROI per dollar.[^2]

The ceiling is real. Channels saturate. CPMs rise. Competitors bid up the same audiences. The attention game is linear by design. You pay for reach. You stop paying, you stop reaching. There is no flywheel.

Most teams allocate roughly 50-60% of growth effort here. It feels productive. The dashboards move. But you're renting an audience, not building one.

Game 2: Persuasion

Persuasion is the game of converting attention to action. Positioning, copy, objections, social proof, trust signals. You've bought the click. Now you have to earn the decision.

The metrics live at every stage of the funnel: landing page CVR, signup completion rate, activation rate, trial-to-paid. Each drop-off point is a persuasion failure. Either the value proposition isn't clear, the trust isn't there, or the ask is mistimed.

The ceiling here is usually an awareness mismatch or a trust deficit. A visitor arrives from a retargeting ad with one expectation and hits a page built for a cold prospect. The message breaks. They leave.

Persuasion work compounds modestly. The median SaaS landing page converts at 3-5%, with top quartile at 8%+.[^1] A page that converts at 8% instead of 5% keeps working after you close your laptop. But it doesn't compound the way systems do.

Most teams spend 20-30% of growth effort here. It should be closer to 40%.

Game 3: Systems

Systems is the game of building machines that compound. Experiment infrastructure, owned audiences, SEO, automation, signup flow quality, data pipelines that feed decisions.

The metrics are different: experiment velocity, organic traffic percentage, email list growth rate, flywheel speed. These don't appear in a weekly paid channel report. That's part of why teams underinvest here.

The time horizon is also different. A signup flow improvement ships in days but pays off over months as every subsequent campaign runs through a tighter funnel. A 5% reduction in signup drop-off cuts CAC across every channel simultaneously. Paid, organic, referral - all of them benefit from the same fix.

That's the unique advantage of growth engineering over pure marketing. Ad spend is linear. Eng work on signup flows, onboarding, and SEO compounds.

The ceiling in Game 3 is team capacity and growth debt. Not market dynamics. You control it.

Why teams get the priority stack backwards

Most teams run this order: attention first, persuasion when there's time, systems never.

The right order is the reverse. Systems first. Then persuasion. Then attention.

The reason is simple: attention poured into a broken funnel is wasted money. If your signup flow loses 40% of visitors and your activation rate is poor, doubling your ad budget doubles your waste.

There's a clean diagnostic for this. Ask yourself: if you 10x'd your traffic tomorrow, would your funnel handle it? If the answer is no - or if you hesitate - your systems game is underdeveloped. Fix that before spending another euro on acquisition.

Measurement is the prerequisite for everything. You can't optimize a funnel you can't see. Before you play any of the three games well, you need to know exactly where attention drops off, where persuasion fails, and which systems are actually compounding versus coasting.

What good allocation looks like

High-performing growth teams distribute effort roughly like this:

~40% on activation and signup. This is the highest-leverage work per engineering hour. A tighter onboarding flow benefits every channel simultaneously. The credit is diffuse - nobody gets a dashboard showing "activation improvement drove 200 new paid accounts this week" - but the impact is real and durable.

~30% on organic systems. Content, SEO, email lists, referral loops. These are slow to build and hard to attribute, which is why underfunded teams cut them first. They're also the only assets that survive a channel shutdown.

~20% on paid optimization. Not growth - optimization. Better creative, better targeting, better landing pages for paid traffic. Not more spend. More efficiency on existing spend.

~10% on exploration. New channels, new formats, experiments with no guaranteed return. This is how you find the next 40% allocation before you need it.

The anti-priorities

Some work is common, legible, and mostly wasted.

Redesigning pages without conversion data. You're redecorating a room without knowing why guests leave. The new design might lift CVR. It might hurt it. You won't know because you didn't baseline first.

A/B testing button colors when the value proposition is unclear. Color tests make sense when everything else is dialed in. They're noise when the page doesn't explain what you do or why it matters.

Launching new channels before existing ones are optimized. Every new channel is a new funnel to build, monitor, and maintain. Add channels only when current ones are performing and you have capacity to do the new one properly.

Copying big-company playbooks at small-company scale. Meta's growth team has 50 engineers and a decade of infrastructure. Their tactics assume experiment velocity and data quality you don't have yet. Build for your stage, not theirs.

Playing all three

The teams that compound fastest play all three games simultaneously - but in the right proportion and the right sequence.

They instrument everything first. Then they tighten the funnel. Then they build systems that make future improvements easier. Then they bring in traffic to a machine that's ready for it.

That sequence feels slow in the first quarter. By the end of year one, the teams that rushed straight to attention are resetting their budgets while the systematic teams are watching their CAC fall without increasing spend.

Three games. Most teams are only playing one.

[^1]: Unbounce, "Conversion Benchmark Report", 2021. [^2]: ProfitWell, "The SaaS Spend Stack", 2022.