Furniture Ecommerce Conversion Case Study in Pakistan

The Client

A Rawalpindi-based custom solid-wood furniture brand with an in-house workshop and a single experience studio in the twin cities. The business had been running for just over a decade, family-founded, and had built its reputation on made-to-order pieces — sofas, beds, dining sets, wardrobes, and occasional furniture — crafted from seasoned sheesham, rosewood, and oak with a range of upholstery options. Roughly seventy percent of what they sold was made-to-order to a buyer’s specifications; the remaining thirty percent was stocked fast-moving pieces.

Online revenue sat near PKR 90 million a year against roughly 70,000 sessions a month, with the workshop and studio still the larger revenue channel. Traffic came from Meta catalog ads, Google Shopping, and a growing organic base, and it was genuinely qualified — buyers were arriving on specific made-to-order products with intent. The blended online order value hovered around PKR 64,000, with custom dining sets and sofa configurations frequently clearing PKR 180,000, firmly in high-consideration territory.

The team’s operational pain lived in a specific place. Because every custom order had been quoted and specified over WhatsApp, the website acted as a gallery rather than a store. A buyer would land on a sofa, see a single size and a single price, message the brand to ask whether it could be made three feet longer in a different wood with a different fabric, wait hours or days for a reply and a quote, and more often than not drift to a competitor or defer the purchase. The brand was paying to send qualified buyers into a manual, asynchronous queue it then lost half of. They came to WeProms Digital with a direct question: can the product page let a buyer actually configure their piece, see the price and the wait, and commit — without the WhatsApp loop?

The Problem: A Product Page That Could Not Sell What the Brand Actually Made

Four issues were throttling the funnel at the product page.

1. Add-to-cart was stuck at 3.4% on a catalog that was seventy percent custom. Sessions were growing, but the share that produced an add-to-cart event was flat. The PDP showed one fixed size and one fixed price for products that were, in reality, built to order. Buyers who wanted anything other than the default had no path on the page — only a WhatsApp number.

2. No live pricing or lead time for custom specifications. A buyer asking for a larger sofa, a different finish, or an upgraded fabric had no idea what it would cost or how long it would take until a human replied. In a category where price and timeline are the two facts that determine commitment, asking the buyer to wait for both was structurally fatal. Most did not wait.

3. Checkout captured no specifications. When a buyer did add the default product to cart and check out, the form collected shipping details but none of the custom choices they actually cared about — wood, fabric, dimensions, finish. Those were collected afterward, by phone or message, which produced measurement errors, rework, and a cancellation rate above nine percent on made-to-order items. The brand was losing margin to a spec-capture process that checkout should have handled.

4. Full upfront payment on an unbuilt item. The store demanded the entire order value before a single piece of wood was cut, on a product the buyer had never seen built. For a PKR 180,000 dining set, that was a wall. Pakistani furniture buyers are accustomed, in the studio, to paying a token to begin work and the balance on inspection at delivery — a trust pattern the website ignored entirely.

Phase 1 — Funnel Diagnosis and Configurator Data Audit (Weeks 1-2)

We began by quantifying exactly where the funnel leaked and by assembling the data the configurator would depend on, because a configurator is only as honest as the pricing and lead-time logic behind it.

Funnel and behavior analysis. GA4 ecommerce tracking was rebuilt and validated server-side so purchase and add-to-cart data were clean and deduplicated. Heatmaps and session replay went across the top thirty made-to-order PDPs and the full checkout. The picture was consistent: strong session-to-PDP progression, a sharp falloff at add-to-cart, and a clear behavioral signal — buyers were selecting the WhatsApp button on the PDP far more often than the add-to-cart button, then disappearing. The website was actively routing committed buyers out of the funnel.

Configurator data audit. The configurator needed three data layers wired to the workshop: a dimension matrix (standard sizes plus custom ranges per product type), a finish and material catalog with real cost deltas, and a lead-time engine keyed to workshop capacity and material availability. We audited the catalog and found that pricing for custom specs lived in the founder’s head and a spreadsheet, finish options were undocumented online, and lead times were quoted case by case. We codified all three into structured data the PDP could read.

Data layer	Before	After
Custom dimensions	Quoted per enquiry on WhatsApp	Dimension matrix with min/max ranges per product type
Finishes and fabrics	Undocumented online, shown on request	Photographed swatch catalog with price deltas
Lead time	”Two to four weeks” verbal estimate	Engine keyed to workshop capacity and material stock
Pricing	One fixed price per product	Live price recompute from selections

Baseline locked. Add-to-cart 3.4%, store CVR 1.2%, checkout completion 41%, blended AOV PKR 64,000, mobile add-to-cart 2.7%, custom-order cancellation rate 9.2%. We also captured WhatsApp-click rate on the PDP as a tracked metric, since it was the leak the configurator was designed to close.

Phase 2 — Configurator PDP for Made-to-Order Buying (Weeks 3-5)

The product page rebuild was the center of the engagement. We treated the PDP as the place where a buyer specifies and commits to a custom piece, and we built it to answer the questions a made-to-order buyer actually has.

Inline configurator with live pricing. Each configurable PDP gained an inline configurator: dimension selectors (standard sizes plus a custom range with clear minimum and maximum buildable limits per product type), finish and material options drawn from the audited swatch catalog, and upholstery fabric choices. As the buyer selected, the price recomputed live from the pricing data layer, and the lead-time engine updated the delivery window on the same screen. The single biggest source of drop-off — not knowing what a custom piece would cost or how long it would take — was answered instantly on the page. Buyers no longer had to leave the funnel to get a quote; the funnel became the quote.

Craftsmanship and material trust. Custom furniture bought unseen online is a trust problem before it is a usability problem. We added a craftsmanship layer: material origin and seasoning notes, joinery and construction detail, warranty terms, and a workshop-process timeline showing how a piece moves from order to kiln-drying to assembly to finish. Real customer delivery photographs accompanied each major category. For a category where the buyer is committing real money to wood they have not touched, this layer did the work that a showroom walkthrough does.

Measurement confidence. Custom furniture has a failure mode the configurator could create: a buyer configures a piece that will not fit through their door or in their room. We added a printable measurement guide and a room-fit reference, and — critically — flagged oversized configurations with a note prompting the buyer to confirm access and room dimensions. This protected the downstream cancellation rate, because a configurator that lets a buyer specify an unbuildable or undeliverable piece would lift add-to-cart while inflating returns and disputes.

Swatch photography as a configurator asset. A configurator that updates the price but shows a flat colour block for a finish is asking the buyer to trust a swatch they cannot read. We ran a focused production to photograph each wood finish and upholstery fabric as a real material sample — grain, sheen, stitching, and edge — so the configurator’s finish and fabric selectors rendered actual product photography rather than placeholder colour chips. For a category bought on appearance, this was the difference between a configurator that felt credible and one that felt like a form. It was the single highest-leverage content investment of the project, because every configuration a buyer built was assembled from imagery they could actually evaluate.

Mobile-first configuration. With more than seventy percent of traffic on mobile, the configurator was built thumb-first: large tap targets, a sticky price-and-lead-time summary that updated as selections changed, swipeable finish and fabric galleries, and a sticky add-to-cart bar. Configuring a custom sofa on a phone had to feel as natural as browsing one.

Phase 3 — Spec-Confirmation Checkout and Payment Path (Weeks 4-8)

With the configurator turning browsers into committers, we rebuilt checkout to capture exactly what the buyer had specified and to remove the payment wall that had been killing big-ticket orders.

Spec confirmation before payment. The first checkout step was a spec-confirmation page summarizing the buyer’s exact configuration — dimensions, wood, finish, fabric, lead time, and total — and asking them to confirm or adjust before entering shipping or payment. This single step did two things at once: it gave the buyer a final moment of clarity on a high-consideration purchase, and it captured clean specifications into the order record, eliminating the manual spec-collection that had been driving rework and cancellations. The buyer’s selections became the build sheet.

Deposit-on-order, balance on delivery. We replaced full upfront payment with a deposit-on-order structure — thirty percent to begin workshop production, with the balance due on delivery and inspection. For a PKR 180,000 dining set, lowering the commitment barrier from the full amount to a token transformed what buyers would commit to. The balance-on-delivery-with-inspection term mirrored exactly how the brand’s studio had always closed sales: pay a token, watch it being built, settle when it arrives and you have seen it.

Slot booking and COD ordering. Delivery and assembly slot booking flowed through checkout, giving the workshop a concrete date to plan against and giving the buyer certainty on when their custom piece would arrive. Payment methods were reordered with cash on delivery prominent for the balance, alongside bank transfer, reflecting how most Pakistani furniture buyers prefer to settle the bulk of a large order — on inspection, in person.

Guest checkout. Account creation was made optional, offered only after order completion. For a category a household buys a few times a year, forcing a permanent account was friction with no payoff.

Phase 4 — Optimize, Validate, and Measure (Weeks 8-12)

With the rebuild live, we ran a focused set of refinements to validate and compound the gains: configurator default selections (pre-selecting the most popular size and finish so the buyer started from a sensible baseline), lead-time copy testing, the placement of the deposit terms on mobile, and the visibility of the spec-confirmation step.

Guardrails were tracked throughout to confirm the add-to-cart lift was not coming at the cost of order quality. Custom-order cancellation rate, workshop rework rate, and time-to-first-contact all improved rather than degraded — a critical check, because a configurator that lets buyers specify unbuildable pieces would have lifted add-to-cart while destroying margin downstream. The measurement-confidence flags and the spec-confirmation step were the specific mechanisms that protected this: buyers committed to pieces they had actually understood and that the workshop could actually build. The deposit option nudged average order value upward, because it lowered the barrier to selecting larger sizes and premium finishes — buyers who would not have paid the full amount for an upgraded configuration upfront committed to it happily under a deposit-then-balance structure.

By the 90-day mark, add-to-cart had moved from 3.4% to 4.96%, and the higher-quality, fully-specified add-to-cart cohort progressed further through checkout than the old, under-specified one.

Final Results at 90 Days

Metric	Before	After	Change
Add-to-cart rate	3.4%	4.96%	+46%
Store conversion rate	1.2%	1.7%	+42%
Configurator-to-add-to-cart	—	58%	New capability
Checkout completion	41%	53%	+29%
Mobile add-to-cart rate	2.7%	4.0%	+48%
Average order value	PKR 64,000	PKR 71,000	+11%
Custom-order cancellation rate	9.2%	3.6%	-61%

These figures are illustrative of the outcome shape a Pakistani custom-furniture or made-to-order retailer can use to sanity-check fit, not an audited third-party result. The add-to-cart lift sits comfortably inside the realistic band for a high-consideration, configurable category starting from a non-configurable PDP.

What Made This Work

1. The configurator closed the WhatsApp leak. Moving custom specification and pricing out of an asynchronous messaging queue and onto the page meant a buyer could configure, price, and commit in a single session. The website stopped handing committed buyers to a manual process it then lost.
2. Live pricing and lead time answered the two questions that govern commitment. In made-to-order furniture, cost and wait are the facts buyers need. Surfacing both instantly, against selections, removed the uncertainty that had been ending most enquiries.
3. Spec capture in checkout killed the rework tax. Capturing exact specifications as the build sheet, at confirmation, eliminated the manual spec-collection that had driven a nine-percent cancellation rate. The conversion lift compounded because the orders that landed were clean and buildable.
4. Deposit-on-order matched local buying reality. Lowering the barrier from full payment to a token let buyers commit to higher-value configurations, lifting both conversion and average order value simultaneously.
5. Trust content did the showroom’s work. Material origin, construction detail, process timeline, and real delivery photographs gave buyers the confidence to commit to wood they had not touched — the online equivalent of a studio walkthrough.
6. Mobile-first configuration met buyers where they were. With most traffic on mobile, a thumb-first configurator with a sticky price summary addressed the device where the funnel had been steepest.

What Teams Can Apply

For Pakistani retailers selling made-to-order or configurable goods online:

1. Move configuration and pricing onto the page. If your buyers routinely message you to ask what a custom version costs, your PDP is leaking committed intent into a manual queue. An inline configurator with live pricing turns that queue into conversions.
2. Make lead time a feature, not a hidden risk. Buyers abandon when they do not know the wait. Surface a workshop-capacity-aware timeline alongside selections and commitment goes up, not down — uncertainty, not the wait itself, is what kills the sale.
3. Capture specifications inside checkout. If your team re-collects specs after the order, you are paying for that gap in rework and cancellations. The buyer’s selections should become the build sheet, confirmed before payment.
4. Offer deposit-then-balance for big-ticket orders. A token-to-start, balance-on-inspection structure matches how Pakistani buyers already fund large purchases and lifts average order value at the same time as conversion.
5. Guard against the configurator’s own failure mode. Let buyers configure freely, but flag oversized or access-constrained pieces and confirm room dimensions. A configurator that permits unbuildable specs will inflate returns and erase the margin the conversion lift created.

WeProms Digital has applied this ecommerce conversion optimization framework across Pakistani retailers in custom furniture, mattresses, modular kitchens and wardrobes, and upholstery. The conversion-focused website design that underpins the configurator PDP and spec-confirmation checkout is adapted per category and order value, but the configure-first, capture-specs-in-checkout, deposit-flexible approach stays consistent. For made-to-order furniture and home-goods brands specifically, the configurator and local-payment patterns map directly onto the furniture store industry context we work in most.