How I Think | Jahnvi Raichura

Principles

Six rules I work by

Start with the decision, not the data

Before building a dashboard, I ask: what decision will someone make with this, and how often?

StyleSignal: every style ends in one of four calls: launch, fix, delay or drop.

Trust beats sophistication

A model people don't trust doesn't get used. I keep logic simple and explain every recommendation.

StyleSignal: I chose explained scoring rules over an AI model, and weekly usage reached 86%.

Fix the cheapest way first

The first fix that comes to mind is often the most expensive. I look for options that use what we already have.

RestockIQ: transfers before reorders fixed 38% of stock-outs without any new stock.

Fewer, better choices

Every extra option adds cost, inventory risk and confusion for the customer.

Dew Daily: launching 5 essentials instead of 10 kept every product under $25 and won the retailer.

Prove it small, then scale

A pilot with a control group beats a big launch based on hope.

All three: every product launched through a pilot or an A/B test before rollout.

Design for the person doing the work

The best tool fits into someone's Monday morning, not into a new process they have to learn.

RestockIQ: one ranked list of the 10 things to fix today, not 400 rows.

My process

From problem to measured impact

01 · DiscoverFind the real problem

Stakeholder interviews, data audits, shadowing and time on the store floor.

02 · DefineWrite it down

A PRD with goals, users, scope, user stories and success metrics agreed up front.

03 · PrioritizeChoose what matters

Score options on impact, effort and risk. Be explicit about what's out of scope.

04 · DeliverLaunch with control

Milestones, owners, go/no-go gates and a risk log, so launch has no surprises.

05 · LearnMeasure and iterate

Pilot vs. control, KPIs against targets, then a Now/Next/Later roadmap.

How I measure success

Metrics I hold myself to

Outcome

Did the business move?

Revenue, availability, on-time launches, margin.

Adoption

Do people use it?

Weekly active users, tool usage, repeat behaviour.

Efficiency

Did it save effort?

Hours saved, fewer meetings, faster decisions.

Customer

Is it better for them?

Fewer returns, bigger baskets, more repeat buyers.

Good products come from clear decisions, not more features.