← All posts

Growth Hacking Tools Aren't the Lever. Growth Loops Are

Professional header image for educational tutorial: Growth Hacking Tools Aren't the Lever. Growth Loops Are

I have sat in too many meetings where the first question about growth hacking was "which tool should we buy?" Not "which loop is broken." Not "what does our funnel actually look like." Just which dashboard, which add-on, which shiny new AI feature nobody has time to learn.

I think the question keeps failing because tools do not create growth. Loops do. So let me reframe the buying decision as a stack decision: map the loop you are actually trying to close, buy the minimum instrumentation that measures it, and add nothing new until a measured bottleneck demands it. Otherwise you pay for a dashboard nobody acts on.

I will cover why the which tool question keeps failing, what a growth loop is and how to pick the one worth closing, the minimum instrumentation each of the four loops needs, a worked example instrumenting a SaaS trial activation loop, who should own the loop, and the primary weekly user test I run before adding anything new. By the end you will have a plan for this week that does not involve buying another tool.

Why the Which Tool Should We Buy Question Keeps Failing

I have paid for tools nobody opened more than twice and built dashboards nobody acted on. The purchase felt like progress; the renewal was harder to justify.

That pattern is not personal. Gartner data reported by Chief Martec shows marketers used 58% of martech capabilities in 2020, 42% in 2022 and 33% in 2023, falling every year while spend climbed. Statista figures cited by MarTech put spend at $15.3 billion in 2020 and $23.6 billion in 2023. Martech spend rose from $15.3bn in 2020 to $23.6bn in 2023, a rise closer to 54% than the 35% some outlets reported.

Gartner puts 25.4% of the marketing budget into technology, so a wrong purchase is a real cost, not an experiment.

MarTech reports that Gartner names three causes of low utilization: skills, governance, and stack sprawl. None is "we bought from the wrong vendor." A Gartner press release dated October 29, 2025 puts 45% of martech leaders saying vendor-offered AI agents fail to meet expected business performance. Headline figure only; the body was not available to me.

Purchasing converts an ambiguous growth problem into a completed task, which is easier than naming the loop you are failing to close.

What a Growth Loop Is and How to Pick the One Worth Closing

A growth loop is a closed system where the output of one cycle feeds the input of the next. A funnel ends at a conversion. A loop repeats, and the repeat is what compounds.

Growth hacking as a search term mostly surfaces tactics. That is the trap: a tactic counts only when it is bolted to a loop you can measure. Otherwise it is a stunt with a screenshot.

Pick the loop first by asking where a 20% improvement changes your model most, not where the work is most fun. Usually that is not the loop your team is excited about. If you sell into bigger accounts, that math shifts again.

State your loop in one sentence before you spend: trial starts convert to activated accounts, activated accounts invite teammates, invites produce more trial starts.

One loop at a time, one dashboard a named person reads weekly. In my experience, two dashboards means zero decisions.

On sourcing: published research maps utilization and spend, and Gartner points people at vendor evaluation tools. The loop-first selection framework is mine, built from first principles. Treat it as a position, not a citation.

The Four Loops and the Minimum Instrumentation Each One Needs

With your loop chosen, the next question is what to actually measure. Minimum instrumentation means the smallest set of events plus one dashboard that tells you whether the loop is turning. Not the fullest, not the most impressive. The smallest.

The test for every metric is simple: name the decision it changes. If you cannot name it, you do not need the metric. Useful tests need well-defined metrics that reflect real business objectives, plus instrumentation scoped to the decision.

So each loop gets three numbers: one primary, one secondary, one guardrail. Three numbers, not a forty-tile dashboard.

Anything producing a report nobody acts on is a cost center wearing an analytics badge. Work each loop against this template before you look at a pricing page.

Acquisition: Reach, Qualified Visit, Signup, Next Round of Reach

Acquisition usually gets picked first. It closes when reach becomes a qualified visit, which becomes a signup or first order, then produces the next round of reach. Referral is the strongest version of that last step. An HBR study of over 10 million consumers found about 20% of new customers come through referrals, and they account for 72% of profits, so I treat referral as a first-class metric.

My minimum events: source of first touch, cost per qualified visit, signup rate, and one attribution view I actually trust. One view, not five.

Guardrail: cost per activated user, not cost per lead, so you cannot buy volume that never turns on.

Work this loop when you have demand you cannot capture. If people arrive and fail to activate, you are in the wrong loop. Skip multi-touch attribution suites, third-party intent data, and account-based platforms until a measured bottleneck demands them.

Activation: Signup, First Value Moment, Habit Signal, Expansion

The shape runs signup, first meaningful action, habit signal, invites or expansion. Minimum instrumentation stays small: one written activation definition, time to first value, and the share of signups who hit it. Keep that definition to one event pair. If it needs six events, it is a description of your product, not a bottleneck.

Your guardrail is a day-7 or day-14 return rate, so activation does not collapse into a finished onboarding checklist nobody acts on. Time to first value is measured from signup to the first observable instance of the validated activation event.

This is the loop I see left unmeasured most often while teams buy acquisition tools, which is how traffic rises and revenue stays flat. If I am comparing activation rates across variants, I run the paired t test instead of eyeballing the lift.

Monetisation: Activation, Price Test, Expansion, Revenue Into Acquisition

Once activation holds steady, monetization is the loop to close. Its shape: activation feeds a paywall or price test, that feeds expansion or repeat purchase, and that revenue funds more of the same acquisition. Plan mix sits at the center of it, so know what the marketing mix actually looks like for SaaS and ecommerce before you test a new tier.

Instrument four events: plan mix, conversion to paid, expansion revenue, and involuntary churn from failed payments. Failed payments land in the same churn figure as cancellations, so a payment problem can look like a pricing problem. Recurly's network data puts voluntary churn at 2.34% and involuntary at 1.25%; only the second is fixable with retries and card updaters.

My guardrails: revenue per activated user plus payback period. They stop a discount campaign from buying volume at economics you would never defend on a call.

Skip CPQ, revenue intelligence, and billing experiment suites until a measured bottleneck demands them.

Retention: Usage, Value Moment, Renewal, Referral

Once activation is steady and pricing has stopped leaking, I look at the loop everyone skips. Usage feeds the value moment, renewal or repeat purchase follows, and referral then lowers the cost of the next acquisition round. I measure three things: cohort retention by signup week, frequency of the core action, and churn reasons captured at the moment someone cancels. Measure retention across activated users only, because never-activated accounts flatter the number and hide the real leak. I keep in mind that some secondary users legitimately touch a tool rarely, raw capability scores can mask that. I find retention boring and slow, and I think it is the highest-yield loop on the list, which is exactly why teams skip it.

Worked Example: Instrumenting a SaaS Trial Activation Loop

Sourcing, plainly: no source I found defines minimum instrumentation for a single loop, so this is first principles. Your activation event has to come from your product. Mine is not yours.

The scenario: a self-serve SaaS with a 14-day trial, a few hundred trials a month, and an 11-tool stack. Nobody can say why trial-to-paid sits where it sits.

Step 1. Define activation as an event pair. Here, a created first project plus one invited teammate, both inside 72 hours, a deliberate choice rather than a standard.

Step 2. Instrument only what that needs: trial start, project created, teammate invited, conversion to paid. Four events, not fourteen.

Step 3. One dashboard with four numbers: trials, activation rate, time to activation, trial-to-paid. Give it one owner and one weekly review slot on the calendar.

Step 4. Leave the other eight tools alone for a full cycle. Do not rip them out mid-experiment and do not add anything new to them.

Step 5. Read the dashboard before buying anything. If time to activation is the leak, the next purchase is onboarding instrumentation, not another lead source.

Step 6. Write the bottleneck in one sentence and date it. A bottleneck you have not written down is an opinion, and opinions are how you end up with dashboards nobody opens.

Three or four events beat a forty-tile dashboard, and the measured bottleneck decides the next tool. This is not a shortcut. The loop mapping is the work, and buying the tool first is how people avoid it.

Growth Agency or In-House: Who Should Own the Loop

So who should run that loop? The question is not whether to hire a growth agency, but which loop you want one to own and what you will measure them on. Agencies are good at execution capacity inside a loop you have already defined and instrumented, and bad at defining your loop for you. If the agency picks the metric and reports on it, you have outsourced the diagnosis and kept the blame. Ask which loop they would start with, and why.

Three questions before you sign: which loop first, which single event defines success, and what shows up on the weekly dashboard. Vague answers usually mean a tool-first agency with a stack recommendation attached.

Keep tool access, event definitions and dashboards in your own accounts. Instrumentation living only inside an agency disappears with the engagement. Ask for a named primary user inside the agency for every tool they add, and a monthly line on what that person did with it.

Full disclosure: dmgoi.com covers SaaS, tech and ecommerce growth tactics. I am not selling agency services, so read this as a buyer checklist, not a pitch.

When to Add the Next Tool: The Primary Weekly User Test

Once the loop has an owner, stop asking what to buy next. My rule: add nothing until a measured bottleneck on that loop demands it, and write the bottleneck down before you open a pricing page. If you can't name it in a sentence, you don't need the tool.

Then run the primary weekly user test: every tool needs one named person who opens it at least weekly. If there is no name, I do not renew. It's the simplest stack audit I know.

I keep in mind that some secondary users legitimately touch a tool rarely. The standard is one primary user with weekly usage, not heavy usage from everybody.

Diagnose before you buy. MarTech reports that Gartner traces underuse to skills, governance, and stack sprawl, and Domenic Colasante, CEO of 2X, has noted that vendors offer certification programs rather than real training programs, meaning teams often never develop working fluency with the tools they already pay for. Budget hours, not just licenses.

One event definition, one dashboard, one owner often shows the answer was never a tool.

Then ask each quarter which tool had no primary user open it weekly. Those contracts go first.

What I Would Do This Week Instead of Buying Another Tool

So here is my actual week.

First, I pick one loop: acquisition, activation, monetization or retention. I write it as a single sentence and name the stage I believe is leaking.

Second, I define the one event that proves the loop is turning, then instrument only what measures it. Four events, one dashboard, one owner, one weekly review.

Third, and here the primary weekly user test applies directly, I give every existing tool a named primary user who opens it weekly and put anything without a name on a cancellation list.

Fourth, I wait. A measured bottleneck earns the next purchase, nothing else.

Growth hacking tools never created my growth. The loop I finally mapped and measured did.

Conclusion

Growth hacking tools were never the lever for me, and every week reinforces it. Pick one loop and name the leaking stage. Define the single event that proves the loop is turning. Give every tool a named owner who opens it weekly. Then wait, because only a measured bottleneck earns a purchase.

The answer was rarely in the next tool. It was in the loop I finally mapped and measured. Start with the loop you already have, and let the numbers tell you what to buy.