06
Trust: the foundation every other driver depends on
Low trust drags App Pulse down across every other driver. High trust on its own does not lift it.
Trust is the foundation every other driver compounds on. Low trust drags products below the market. High trust on its own does not lift them above. It is the smallest of the four drivers by share of App Pulse, and the most consequential by mechanic. Chapter 5 looked at what users see. This chapter looks at what they trust: whether the product acts in the user’s interest, and whether it handles data responsibly.
5min
Scroll down
Low trust drags App Pulse down across every other driver
Low trust pulls a product below the market, whatever else the product does well. High trust on its own does not lift the product above. Trust accounts for 21 percent of App Pulse, the smallest share of the four drivers, and the one that does the most damage when it is missing.
Eight of the top 10 apps on “handles data responsibly” are Danish, led by Sydbank. The Netherlands is represented by Rabobank. Sweden by Skatteverket, the tax agency. Trust does not travel across markets. The signals that earn it in Denmark are not the signals that earn it in Sweden. A product team building cross-market needs to understand the local norms before designing the trust layer.
The pattern goes further than country leadership. Rank every app in the study by trust and 28 of the top 30 are built for a single market. Zero of them appear in all three markets at the top. The bottom of the same ranking is dominated by global products: ChatGPT, Facebook, Instagram, TikTok, X, Temu.
3.8
is the average trust score for single-market apps.
3.51
is the average for apps operating across all three markets.
A global brand can reach further than a local one. It cannot inherit the trust the local brand has earned. For a global product entering a new market, the trust ceiling is local work, not a brand-asset transfer.
The top apps do not just avoid low trust. They lead on it. “Acts in my interest” shows the widest gap between the best and weakest apps of any statement in the study. Trust is not only the foundation you protect. It is where leaders create distance the rest of the market has not closed.
Trust score by category
Provided by framna
Provided by framna
Where does your product ask for trust without earning it first?
Innovation can be perceived without trust. Retention cannot
Among the most innovative apps in the study, trust scores split sharply. ChatGPT and Gemini lead on innovation while scoring well below market on trust. Lunar and Tibber pair both. Innovation can be perceived in the absence of trust. What trust gates is retention. For a product adding AI features, innovation may land first; trust decides whether users stay.
For a team solving user frictions in the next 12 months, this defines the sequencing. Innovation does not require an AI feature. It requires solving a real problem in a way users did not expect. When the solution does involve AI, the trust investment ships in the same release: visible data handling, an explicit reason for every recommendation, a one-tap way to override the system. Teams that build trust alongside the capability keep the users it brings in.
Claude is the counter-example with an above-market-average score on trust. The product makes safety, reasoning, and data handling visible in the interface. The trust deficit is a design choice, not a property of the category. But trust is not only a product-level design decision. Brand reputation, track record, and public positioning all shape how much goodwill a user brings to the first session.
Top-decile innovation apps and their trust scores
ChatGPT and Gemini sit high on innovation, low on trust. Lunar and Tibber pair innovation with strong trust. Innovation perception is not gated by trust. Retention is.
Trust scores for top-decile innovation apps, with cross-reference to NPS and retention indicators.
Provided by framna
Provided by framna
If your product is shipping AI features in the next 12 months, what is the trust architecture you are building alongside?
The gender gap is not about features. It is about trust
Women rate trust statements consistently higher than men, across every category in the study. On “Handles data responsibly”, women score 3.7 against men’s 3.6. “Acts in my interest” shows the same shape. On technical performance the gap closes to under 0.05. The split is on trust dimensions, not on whether the product works.
The gap is small individually, consistent at scale
On a 1 to 5 scale, a 0.1 gap looks like a rounding error in a dashboard. At population scale, with thousands of respondents per market, the gap is consistent in direction and magnitude across every category in the study.
The pattern is not category-specific
Men enter every product in the study with less goodwill on the two dimensions that decide whether they grant the product permission to act on their behalf. The pattern holds across Finance, Health, News, and Shopping at the same magnitude.
Most teams do not measure it
Most measurement frameworks aggregate gender, which means the gap is invisible to the team. Cut trust metrics by gender, or the product gets optimized for the half of users who arrived more forgiving.
The product decisions that come next (AI features, recommendation engines, personalization) all ask the user to grant trust to a system the user does not see. Half the user base enters that ask from further behind. Most measurement frameworks aggregate gender, which means the gap is invisible to the team. Cut trust metrics by gender, or the product gets optimized for the half of users who arrived more forgiving.
Women and men: driver-level score gaps
Women rate every driver higher than men; the gap widens on trust dimensions.
Weighted average per (statement × gender) across significant app-country pairs. The gap is computed as women minus men.
Provided by framna
Provided by framna
Half the user base enters the product with less goodwill. Which half is measured?
What this means
Finding
Low trust drags App Pulse down across every other driver. High trust on its own does not lift it.
Evidence
Trust accounts for 21 percent of App Pulse, the smallest share of the four drivers. Eight of the top ten trust apps are Danish. Men rate trust 0.1 below women across every category in the study.
Implication
Trust is the foundation every other driver compounds on. Build the trust architecture into the same release as the AI features that need it.
What we did not expect
We expected trust to lead in Finance and Medical, and to be the foundation that lets personalization and AI features land. Two surprises. The foundation effect runs across every category, not just Finance and Medical: low trust drags every other driver down, and high trust on its own does not lift them. The trust signal itself stays local. 28 of the 30 highest-trust apps in the study are built for a single market. The bottom of the ranking is global products. For a local product, trust reads as a structural moat: hard to build, harder to copy, and not transferable when a global brand enters the market.