Harsha Kotthapalli
← All 34 outputs

Measure and maintain · output 31 of 34

Trust metrics: whether the market believes you, measured.

Trust in three layers, strongest first: what people do, what they say when asked one question, and what others lend. Read quarterly; the first reading names the cheapest instrument, which usually costs one message.

Trust metricsQ3 · R1
What people do
What they say
What others lend
0/12
Measured
Pass line · 4 in 10
Nobody has checked whether anyone trusts this. One question to the 60 checks, free.
Quarterly · the Ellis question, never NPS · borrowed items chain to the Claim check
Nobody has checked whether anyone trusts this. One question to the 60 checks, free.

The full output carries the three layers, the twelve items, the one question and the pass line drawn before anyone answers.

See the full output →

The output at a glance · built on WealthKarma

How it gets made

What goes in

Behaviour counts
The one question to users
Regulator, custody, ratings
Vouches and reviews

Behaviour outranks answers, answers outrank badges

What comes out

The trust ledger

The report

The pass line written before the survey

How long

1day

per reading

quarterly

The output, in full

Built on WealthKarma, a retirement app for Gulf expats. Every count is real and every gap is marked as a gap.

Trust metricsWealthKarma · reading 1 · Q3 2026
Trust in three layers, strongest evidence first. Read quarterly; trust moves slower than the funnel.
01 · The ledger at a glance
0/12
Items measured
? / the one question
Never asked
1/5
Borrowed items with anything behind them
02 · Three layers, strongest first
1 · What people do0 of 4 measured
Kept coming back
Brought a friend
Put real numbers in
Employer renewed vs baseline

All four wait on the funnel reading's instruments. Behaviour is the strongest evidence and all of it is dark.

2 · What they say when askednever asked
"How would you feel if you could no longer use this?" · pass line 40% "very disappointed", written now

NPS not used: the 2007 replication found it predicts growth no better than plain satisfaction questions.

3 · What others lend1 of 5 has anything behind it
A named regulator
Custody, stated plainly
Ego name · permission unrecorded
App-store rating, with its count
A finfluencer vouch
44% trust finfluencers; 57% of those would rethink a company one vouches for (Edelman 2026)

Every borrowed item chains to the Claim check. Borrowed trust you cannot back is a liability with a date on it.

03 · The finding
Nobody has checked whether anyone trusts WealthKarma with their retirement. One question to the 60 beta users would check, and it costs nothing.
04 · The cheapest instrument, ready to run
The message · WhatsApp, this week

"Quick one: how would you feel if you could no longer use WealthKarma? Very disappointed · somewhat · not disappointed."

The pass line, drawn before anyone answers
4 in 10 saying 'very disappointed' is fit

Needs 30 answers of the 60 to mean anything. Cost: one message.

05 · What can't be known
The behaviour layer stays dark until the funnel reading's five instruments exist; this ledger reads them, it cannot build them.
One survey is a point, not a trend; the second reading, next quarter, is where the ledger starts to speak.
What the wider market thinks needs a paid research panel. That is a purchase, not a failing, and the ledger says where it begins.

Where it sits

Research

Compete

Position

Message

Launch

Measure

you are here · 31 of 34

Want this for your company

This is one of thirty-four. The diagnostic tells you which ones you're missing.