By Yulia Farber
After a decade of hypergrowth metrics defining success, the companies built to last are those that users actively choose to trust.
For ten years, the dominant logic of the technology industry was simple: grow fast, fix later. Metrics like DAU, MAU, and viral coefficient became the grammar of strategy. Trust — that slow, unquantifiable thing — was treated as a byproduct of scale, not a prerequisite for it. The companies now discovering this assumption was wrong are doing so the hard way.
In 2021, a watched one consumer fintech startup celebrated reaching one million users in 14 months. Their growth team had mastered the dark pattern playbook: forced notification permissions, guilt-trip cancellation flows, pre-ticked upsell boxes. The metrics were extraordinary. Then came the reviews. Then the churn. Then the regulator. Then silence.
They are not an outlier. The 2024 Edelman Trust Barometer found that trust in technology companies has fallen to its lowest point in a decade, with 57% of global consumers reporting they now actively limit their use of digital products they once used daily. The cost of distrust is no longer abstract — it shows up in cohort curves, in support queues, in the silence of users who simply leave and never explain why.
The decade that mistook engagement for loyalty
The growth playbook of 2012–2022 was built on a powerful but flawed assumption: that a user who comes back is a user who trusts you. The metrics seemed to confirm it. Retention rates, session lengths, click-through rates — all positive. The problem was that these numbers measured behaviour, not belief. They could not distinguish between a user who returned because they loved the product and one who returned because they had no better option yet.
Dark patterns accelerated this confusion. A 2023 study by the Norwegian Consumer Council found that over 90% of top-grossing apps deployed at least one manipulative design pattern — from deceptive defaults to artificial urgency. These patterns worked, in the narrow sense: they increased short-term conversion. But they were quietly accumulating a debt that would eventually come due.
That debt is now being called in. Apple’s App Tracking Transparency removed the ability to retarget without consent. The EU’s Digital Markets Act imposed structural constraints on platforms that had built business models on user data extracted without meaningful permission. And users — increasingly aware of how their attention is being harvested — began opting out of the entire category of extractive digital experiences, not just individual products.
Engagement had never been loyalty. It had been captivity with a good UX.
What trust actually means as a business asset
Trust is not a satisfaction score. It is not an NPS of 70 or a five-star rating or a “likely to recommend” survey result. Those measure how a user feels about a past interaction. Trust measures something different: whether users believe you will behave well in situations they have not yet encountered.
Call it Anticipated Goodwill — the credit a company accumulates when users assume, by default, that its choices will be in their favour. It is why people buy Apple products without reading the privacy policy. It is why Stripe founders who leave Stripe still recommend Stripe to their successors. It is why Notion grew almost entirely through word-of-mouth among knowledge workers who trusted a friend’s recommendation more than any advertisement.
Anticipated Goodwill compounds. Each decision a company makes that prioritises user interest over short-term extraction — a transparent pricing page, a genuinely easy cancellation flow, a product update that adds value rather than harvests data — deposits into an account that accrues interest over time. Each manipulative decision makes a withdrawal. And withdrawals, unlike deposits, tend to be irreversible.
Exhibit 1: Two models of growth — how they differ across every dimension of the business
| Dimension | Growth at all costs | Trust-first model |
| Primary metric | MAU / DAU / viral coefficient | Referral cohort LTV, permission grant rate |
| Acquisition model | Paid channels, dark patterns, aggressive defaults | Word-of-mouth, product-led, low CAC
|
| Retention driver | Engineered friction, exit barriers | Genuine value, Anticipated Goodwill |
| Risk profile | Regulatory exposure, churn cliff, brand collapse | Slower early growth, durable competitive position |
| User behaviour signal | Returns because alternatives are limited | Returns and recommends — unprompted |
Source: Author’s analysis
The compounding economics of earned trust
There’s a reason Stripe’s revenue per employee outpaces category averages: their users don’t need to be re-acquired.
Forrester’s 2024 CX Index found that customer-obsessed companies grow revenue 41% faster and profit 49% faster than competitors who don’t. Only 3% of companies currently qualify. The gap between knowing this and acting on it is where trust is won or lost.
Notion hit 20 million users with a marketing team smaller than most companies’ design departments. Linear runs an NPS above 70 in a category where 30 is considered strong. Neither built this through acquisition spend. They built something users wanted to tell other people about.
Trust reduces every cost in the funnel simultaneously. That’s not a soft asset. That’s a structural advantage.
Three signals that reveal whether users actually trust you
Most companies don’t know their real trust level. They know their satisfaction scores. Different thing.
Trust shows up in three places — none of them on a standard growth dashboard.
Voluntary permission: users who trust you grant access they don’t have to grant. Optional data sharing, feature opt-ins, privacy settings left at default. A private decision, made repeatedly, when no one is asking.
Unprompted defence: when something goes wrong, trust-surplus users absorb it and explain it to others. Watch your community forums after an incident. That’s your real trust reading.
And the one that lives in your own data: referral cohort performance. Users who arrived through a recommendation consistently show higher LTV, lower churn, lower support costs than paid cohorts. If the gap is large — you have trust. If it’s narrow, or referral share is negligible, you have a product people use but don’t believe in.
That distinction matters more than any headline metric.
Exhibit 2: Three measurable signals of user trust — and what each reveals
| Signal | How to observe it | What it tells you |
| Voluntary permission | Optional data sharing, feature opt-ins, default settings left unchanged | Users believe your defaults are in their interest |
| Unprompted defence | Community forums, social media threads defending the product during incidents | Users absorb imperfection because the relationship capital is high |
| Referral cohort premium | LTV, churn, and support cost comparison between referral and paid cohorts | Trust converts users into a self-sustaining acquisition engine |
Source: Author’s framework
What this means for how you build
The shift from growth-at-all-costs to trust-first is not a values statement. It is a strategic reorientation, and it begins with measurement. If you are not tracking referral cohort performance, permission grant rates, and unprompted advocacy alongside your standard growth metrics, you are optimising for a model with a shortening shelf life.
It also means making different decisions at the product level — and doing it before the regulator or the churn curve forces you to. When a dark pattern would increase a metric, it is asking you to borrow from your trust account to pay today’s number. Most teams take that trade without pricing the interest rate.
You already sense this. The products you recommend to people you respect are not the ones that manipulated you into converting — they are the ones that gave you something you did not expect and asked for nothing in return. That instinct is not sentiment. It is a signal about what lasts.
Conclusion
The metrics that defined the last decade of technology — MAU, DAU, viral coefficient — measured behaviour. The metrics that will define the next one measure belief: do your users think you will act in their interest, even when they are not watching? Companies that have earned that belief carry an asset no acquisition budget can buy and no competitor can copy quickly. The work is slower. The compounding is real.
Three things worth doing this quarter: audit your product for trust-withdrawing patterns and price them against cohort lifetime value; measure referral cohort performance separately from paid acquisition; and identify your ten longest-tenured users — then ask them what they would lose if you disappeared tomorrow. Their answers are your actual strategy.
Yulia Farber
