Designing for Conversion, Not Screens

Context

A modern product sold through an outdated workflow

DJI is known for frontier-level consumer drone technology and professional-grade tools. D1 Store is their authorised Australian retailer, selling across 11 physical stores. Despite representing a cutting-edge brand, the in-store retail experience had not evolved at the same pace.

Stores were busy and interest was high, but conversion was inconsistent. Customers spoke with staff, explored products, asked for discounts — then left. Sometimes with a paper quote. Sometimes with nothing. Once they walked out, the store lost the thread.

A modern product was being sold through an outdated retail workflow — creating measurable conversion loss once customers left the store.

Research

How sales actually happen on the shop floor

I started on the shop floor. Interviewing customers directly wasn't viable — disrupting live sales wasn't an option. Instead, I focused on staff as high-volume observers of real behaviour. In the Pitt Street Mall flagship, weekend footfall reached 150–200 customers, with each staff member interacting with ~40–50 customers per day.

I shadowed live sales conversations end-to-end: how customers browsed, where staff stepped in, how pricing was communicated, how quotes were created, and what happened when customers hesitated.

The service blueprint exposed a core issue: the store had no digital memory of intent. Every quote was a dead end.

Insights

Three breakdowns in the retail system

Service blueprinting, shop-floor shadowing, and staff interviews surfaced three core breakdowns:

To ground the work in measurement, I tracked only delayed conversions — quotes that converted one or more days after being issued, using paper quotes as the baseline. This avoided inflating results with same-day purchases and ensured each intervention could be evaluated on whether it preserved intent beyond the visit, not just whether it drove immediate sales.

Each intervention was designed to eliminate a possible lever — until the real one remained.

Intervention 1

A broad test to find the conversion lever

The first intervention introduced a customer-facing kiosk as a deliberately broad test, allowing multiple assumptions to be evaluated in parallel. It tested whether customer autonomy, digital intent capture, and embedded accessory prompts would meaningfully improve conversion.

The kiosk allowed customers to browse and compare drones visually, see compatible accessories in context, and generate a digital quote via email — without staff involvement.

Where conversion fell apart

Returning to the shop floor revealed why the kiosk underperformed as a conversion tool. Two distinct customer patterns became clear:

Browsers were price-sensitive and well-researched, wanting quick reassurance on price. The kiosk introduced friction. Staff frequently bypassed it, giving verbal quotes instead. Without paper quotes during the pilot, many of these customers left without any usable reference point.

Buyers intended to purchase in-store and valued staff expertise. For them, the kiosk proved useful as a staff-led visual aid — not as a self-service tool.

The kiosk improved exploration and attachment, but weakened delayed conversion after the visit. Autonomy itself wasn't driving conversion. At that point, I made the call to abandon it as the primary lever and refocus entirely on pricing speed and intent persistence.

Conversion hinged on fast, staff-led pricing reassurance — setting the direction for a staff-led, digitally persistent quoting flow.

Intervention 2

Restoring speed and digital memory

The second intervention shifted focus to the moment where decisions were actually being made: pricing conversations led by staff. A full POS integration was considered and deliberately rejected — the existing system was bloated and tightly coupled, making it high-risk to modify.

Instead, we built a lightweight sidecar: a staff-assisted quoting flow on a tablet, designed to be as fast as a paper quote while creating a durable digital record. Staff manually added products and generated a price instantly. The interface was deliberately constrained — we pivoted from a discovery hierarchy to a search-and-add hierarchy.

Each staff member logged in under their own ID, allowing quotes to be attributed and tracked. The tablet was handed to the customer to enter their email. The quote was generated immediately, emailed, and stored in the system.

Pricing speed was restored and intent finally persisted beyond the visit — creating a stable foundation for further optimisation.

Outcome

A principle proven in sequence

A 14-day pilot at Pitt Street Mall flagship tested each intervention in sequence against paper quotes as the baseline. Testing in sequence was deliberate: each version isolated one variable to prove causality, not correlation. Only delayed conversions — quotes that returned one or more days after issue — were counted, isolating intent persistence from same-day purchases.

Intervention 3 delivered +11 delayed conversions over the 14-day pilot — the only version to outperform paper quotes on the primary metric without regressing on the others.

This project concluded deliberately without replacing the POS — several tempting directions were left on the table, each of which would have added surface area without addressing the core failure point: pricing hesitation at the moment of intent. By narrowing the system to that moment, the design shifted behaviour without asking staff or customers to change how they already acted.

This reinforced a principle I now apply across complex, high-intent systems: meaningful impact comes from identifying where decisions stall, proving the true leverage point, and designing the smallest possible system that removes friction at that moment — and resisting everything else.