Growth marketing vs growth hacking: what is real.

The phrase is younger than most people assume, and it did not start as a synonym for cheap tricks. In 2010, a marketer named Sean Ellis, who had helped scale Dropbox and other early companies, wrote a short post looking for someone to replace himself. He could not find the right title for the role, so he invented one. In that original essay, he defined the new hire this way: "A growth hacker is a person whose true north is growth. Everything they do is scrutinized by its potential impact on scalable growth."

Read that again, because it is almost the opposite of the caricature. The job was not to find one viral stunt. It was to make growth the lens through which every decision passed, including product decisions that traditional marketers never touched. The word "scalable" is doing quiet, heavy lifting. A trick that pops once and cannot be repeated is, by Ellis's own definition, not what he was hiring for.

So the founding idea was sound. What happened next is the familiar story of a good word getting kidnapped.

A useful phrase met a content economy that rewards excitement, and the two were not good for each other. "Growth hacker" sounded like a cheat code, and cheat codes sell. Conference talks and blog posts stripped the discipline out and left the dopamine in. The genre calcified into a list of one-weird-tricks: the unsubscribe page that begged you to stay, the fake scarcity timer, the LinkedIn automation that messaged strangers until the platform banned you. Tactics, detached from any system, presented as the whole of the craft.

The deeper problem is that the tricks have a built-in expiration date, and it is not a coincidence. The investor and writer Andrew Chen named the mechanism years ago in a piece he calls the Law of Shitty Clickthroughs. His evidence is almost funny: the first banner ad ever shown, on HotWired in 1994, had a clickthrough rate around 78 percent. People clicked because they had never seen such a thing. By 2011 the typical banner pulled roughly 0.05 percent. Same ad format, a difference of more than a thousandfold, driven entirely by the audience getting wise.

• Novelty fades. People click the new thing because it is new. Newness is not a renewable resource.
• Competitors copy. The moment a tactic works publicly, everyone runs it, and the channel gets crowded and expensive.
• The audience shifts. Early adopters forgive friction and gimmicks. The mainstream that follows does not.

This is why your hack fizzled, and it is not a personal failing. Every channel and every clever trick decays on this same curve. The skill is not finding the one trick that lasts forever. No such trick exists. The skill is building something that survives the decay.

Strip away the mythology and growth marketing is unglamorously specific. It is the practice of running small, structured experiments across the entire customer lifecycle, then keeping what works and killing what does not, on a loop, forever. HubSpot puts it cleanly in its overview of growth marketing: it is "the practice of running small, strategic experiments to drive sustainable business results," and unlike traditional campaign marketing, it "extends across the full funnel, touching all aspects of the customer lifecycle from acquisition to retention."

Notice the two words that separate it from hacking: "sustainable" and "retention." A growth hacker in the bad sense obsesses over the top of the funnel, the getting-them-in-the-door part, because that is where the screenshot-worthy spikes live. A growth marketer treats acquisition as one chapter of a longer book. Getting someone to sign up means nothing if they leave on Tuesday. The experiment is not "how do we get more clicks." It is "where in this whole journey are we leaking the most value, and what is the cheapest test that might plug it."

That reframe changes everything downstream. It means a growth marketer will happily spend a sprint improving onboarding, or rewriting a confusing pricing page, or fixing the reason customers churn in month two, because those moves compound, while a viral acquisition stunt does not. The work is less of a stunt and more of a daily habit, closer to flossing than to fireworks.

If you only borrow one framework from the growth world, borrow this one, because it forces you to look past acquisition. In 2007, the investor Dave McClure gave a talk laying out five metrics that, together, describe the full life of a customer. The acronym spells AARRR, which is why everyone calls them the Pirate Metrics. Amplitude has a clear breakdown of the AARRR framework, but the five stages are simple enough to keep in your head.

• Acquisition. How people first find you. The part the hackers fixate on, and only one fifth of the picture.
• Activation. Whether a new user hits their first real "aha," the moment they feel the value. Most signups die here, unseen.
• Retention. Whether they come back. This is the load-bearing wall, and we will return to it.
• Referral. Whether happy users bring others, which is the only honest, durable form of "viral."
• Revenue. Whether any of it pays the bills. The number that makes the rest real.

The value of the map is diagnostic. When growth stalls, the instinct is to pour more into acquisition, the top of the funnel, because that lever is loud and familiar. But the leak is usually further down. You do not have a traffic problem. You have an activation problem, or a retention problem, and dumping more strangers into a bucket with a hole in the bottom just wets the floor faster. The framework makes you find the actual hole before you spend a dollar.

Here is the unsexy center of the whole discipline. Acquisition is addition. You add a customer, you get a customer, and if you stop paying, the adding stops. Retention is multiplication. A customer who stays does not just keep paying. They learn your product, refer their friends, and quietly subsidize the price of acquiring the next person. Growth that compounds is almost always growth built on a base that does not leak.

Andrew Chen made this point against his own younger self. In a reflective essay he wrote ten years after the growth hacking boom, he is blunt about a hard limit he learned: you cannot A/B test your way to product/market fit. If the underlying product does not retain people, no amount of clever optimization at the top rescues it. Worse, he notes that aggressive growth tactics often pile up "UX cruft," little dark patterns and nags that goose a metric this week and erode the experience for months. The hack that spiked your signups can be the same hack quietly poisoning the thing that keeps people around.

This is the cleanest way to tell a real growth team from a hack-chaser. Ask what they measure. If the honest answer is some flavor of "new signups," you are talking to someone filling the top of a leaky bucket. If the answer involves how many of last month's users are still here this month, you are talking to someone building an asset.

The difference is easiest to see in an example everyone knows. Dropbox is forever cited as a growth hacking legend, usually for its referral program: give space, get space. But the thing people miss is that the referral loop was not a stunt. It was a repeatable system wired into a product people already wanted to keep using. Every satisfied user had a built-in, low-friction reason to recruit the next one, and the next one had the same reason. The loop fed itself because the product retained. The "hack" only worked because the unglamorous foundation underneath it was solid.

Contrast that with the stunt that makes the rounds and dies. A brand buys a moment, a provocative video, a giveaway, a borrowed meme, and the traffic floods in for forty-eight hours. Then it evaporates, because nothing about the stunt was designed to repeat, and nothing about the product was ready to keep the people it briefly attracted. There was no loop. There was a firework: bright, loud, gone, and impossible to relight without buying a whole new firework.

• A stunt is an event. It has a start and an end, and when it ends, you are back at zero, shopping for the next one.
• A system is a machine. You build it once, then tune it, and it keeps producing while you sleep and while you iterate.

The test is almost embarrassingly simple. Can you run this again next month, and the month after, and will it still work? If yes, you have a system, and systems are what compound. If no, you have a stunt, which is fine for what it is, as long as you never confuse it with a strategy.

None of this is a secret, which is part of why it gets ignored. It is a rhythm, and the rhythm is more demanding than any single hack precisely because it never ends. You pick the one metric that matters most for your stage, usually activation or retention rather than raw traffic, and you make it the thing the whole team watches. Then you run the loop.

• Find the biggest leak. Walk the AARRR funnel and locate the stage bleeding the most value. Fix the worst leak, not the loudest one.
• Write a real hypothesis. "If we do X, then metric Y improves, because Z." If you cannot fill in the "because," you are guessing, not experimenting.
• Run the smallest honest test. The cheapest, fastest version that could actually move the number. Speed of learning beats size of swing.
• Read it without flinching. Most experiments fail. That is not waste, that is the data. Kill the losers fast and bank the lesson.
• Bank the winners and repeat. Make the wins permanent, then go find the next biggest leak. The loop is the strategy.

The uncomfortable truth is that this is less fun than collecting hacks. There is no single screenshot moment, no Monday-morning brag. There is a quiet engine that, run for a year, produces a chart that only goes up while your hack-chasing competitors are still rebuilding after their last firework fizzled. Slower. It stays.

So here is the part the case studies leave out, because it does not make for a thrilling headline. There is no growth hacking technique that compounds on its own. The compounding lives in the system around the technique: a product worth keeping, a funnel you actually measure, a loop you can run again and again, and the patience to let the slope do its slow work. Anyone selling you one trick that fixes everything is selling you a firework and calling it the sun.

The good news is that the boring version is learnable, and most of it you can do yourself with the frameworks above. Pick your one metric. Find your worst leak. Run the loop. Building that kind of repeatable, full-funnel growth engine is a fair amount of what we do at Mining Wells under Growth & GTM, but the honest answer is that you do not need us to begin. You need to stop chasing the next spike long enough to build the slope. The slowest version is the only one that compounds, which is annoying, and also the entire point.