Overview
SME retailers and distributors run their businesses on a fragile stack of spreadsheets, WhatsApp messages, and paper ledgers, inventory in one file, sales in another, and staff performance tracked nowhere at all. SalesSync brings product creation, stock entries, multi-warehouse inventory, point of sale, shelf-life tracking, and sales prediction into a single sales operations platform across three surfaces (POS, admin portal, marketing website), governed by role-based access so owners, administrators, business managers, inventory managers, and point of sale staff each see exactly what their role needs and nothing more. I led product design across a 4-month engagement, starting with field research in warehouses and shops and carrying through to a tested, launched product that cut order processing time by more than half.
An intelligent sales automation platform that unifies product creation, stock entries, inventory, point of sale, and business intelligence into one system, with shelf-life tracking and sales prediction built in from day one.
The Challenge
For SME retailers and distributors, growth is what breaks their tools. A business managing three hundred SKUs across two warehouses was running inventory on conflicting spreadsheets, ringing up sales on a cash register with no link to stock, and reconciling everything by hand at month-end. Stockouts got discovered when a customer complained. Over-ordering tied up cash in dead stock. And owners had no reliable answer to the basic questions: what sold this week, which staff member is actually performing, which products are about to expire.
The design challenge was bringing software-grade rigor to businesses that had never run on operational software, without asking them to behave like enterprises, across three distinct surfaces: a point of sale for front-of-house staff ringing up customers, an admin portal for the inventory and business managers running the operation, and a public website introducing the product to new businesses. If any single one of those surfaces felt harder than the paper it replaced, the whole system would quietly fall apart.
Business Goals
- 1Cut order processing time dramatically by replacing WhatsApp-and-spreadsheet workflows with structured digital ordering
- 2Bring stockout and over-stock incidents down through real-time inventory visibility and reorder alerts
- 3Cut wastage from expired stock through proactive shelf-life tracking and alerts
- 4Give owners self-serve answers to sales, stock, and staff-performance questions, no more waiting on month-end reconciliation
- 5Turn historical sales data into forward-looking predictions owners can actually plan around
- 6Win real adoption at the point of sale itself. The platform fails outright if staff quietly revert to a plain cash register
User Needs
- 1Owners: a live, trustworthy picture of stock, sales, and cash tied up in inventory
- 2Operations managers: process orders, transfers, and receiving without typing the same data in twice
- 3Front-of-house staff: build and complete a customer order at the point of sale, live stock check included, in under two minutes
- 4Everyone: one source of truth that finally ends the argument about which spreadsheet is current
Key Findings
- 1Every business we audited ran at least three conflicting versions of their inventory spreadsheet. The source-of-truth argument was universal
- 2Order errors clustered at transcription points. Paper tickets and a disconnected cash register, re-typed into spreadsheets, produced most of the reconciliation mistakes
- 3Owners made purchasing decisions on gut feel simply because assembling actual sales data took days of reconciliation
- 4Perishable and date-sensitive stock was tracked, if it was tracked at all, on handwritten labels, so expired inventory was routinely discovered only after it had already gone to waste
- 5Front-of-house staff promised customers stock they had no way to verify without leaving the till, which damaged customer trust the moment an item rang up unavailable
- 6Existing tools, enterprise ERPs on one end and generic POS software on the other, failed these businesses on complexity and price. The SME middle sat underserved
User Voices
"I have three spreadsheets and I promise you, all three tell me a different number for the same product."
"My till says we have stock. My shelf says we do not. I find out when the customer is standing right there."
"Ask me what sold best last month and I will tell you next month, after we finish counting."
User Personas
Affinity Mapping · Key Themes
One Source of Truth
- 1Conflicting spreadsheets turned every number into something up for debate
- 2Trust in the system depends entirely on stock figures matching what's actually on the shelf
- 3Recording a movement once, at the source, kills reconciliation work outright
The Transcription Tax
- 1Most order errors turned out to be re-typing errors, not judgment calls gone wrong
- 2Structured capture at the point of origin eliminates that whole error class
- 3The ops manager's real job had quietly become acting as a human API between tools
Decisions in the Dark
- 1Owners bought on gut feel because real data always arrived weeks too late
- 2Visibility into slow movers and dead stock directly frees up working capital
- 3Managing reps without numbers breeds mistrust on both sides of the desk
Field Reality
- 1Reps work standing up, on mid-range Androids, on data connections that come and go
- 2The tool is competing directly with the till it's replacing, and it has to be nearly as fast
- 3Offline tolerance isn't a nice-to-have here, it's a requirement
Empathy Map · Chuka Eze
User Journey Map
Information Architecture
Dashboard
- 1Sales overview
- 2Stock health
- 3Alerts & reorder points
- 4Rep leaderboard
Inventory
- 1Product catalog & creation
- 2Multi-warehouse stock
- 3Stock entries & receiving
- 4Transfers & adjustments
Shelf Life
- 1Expiration tracking by batch
- 2Threshold configuration
- 3Near-expiry alerts
- 4Wastage reporting
Point of Sale
- 1Product search & scan
- 2Order build & pricing
- 3Payment processing
- 4Receipt generation
Customers & Reps
- 1Customer accounts
- 2Visit & order history
- 3Rep targets & performance
- 4Route activity
Analytics & Prediction
- 1Sales reports
- 2Product movement
- 3Sales prediction
- 4Dead stock analysis
Settings
- 1Users & roles
- 2Warehouses
- 3Pricing & tax
- 4Notifications
User Stories
Approaches Considered
Inventory Tool First
Launch a focused multi-warehouse inventory tracker first, and add ordering and CRM modules in later phases.
Pros
- + Small, shippable scope to start with
- + Attacks the source-of-truth problem head on
Cons
- - Front-of-house staff stay on the old till, so the data entering the system is still second-hand
- - Owner analytics stay thin without a sales flow feeding them
- - The value proposition is too narrow to actually displace spreadsheets
Unified Ops Platform
One platform where inventory, point of sale, and analytics share a single data spine, across a POS surface, an admin portal, and a public website, with the POS feeding structured sales data in right at the source.
Pros
- + Structured capture at origin fixes data quality for every downstream module
- + Owner, ops, and front-of-house each get a purpose-built view of the same underlying truth
- + Analytics finally become trustworthy, because the inputs are
Cons
- - The broadest initial scope of the three options
- - Adoption hinges on winning over front-of-house staff specifically
WhatsApp-Layer Automation
Meet users where they already are: a bot layer that parses WhatsApp orders into a backend, leaving frontline workflows untouched.
Pros
- + Near-zero behavior change required from front-of-house staff
- + A novel, low-friction pitch to make
Cons
- - Parsing paper tickets reliably is a research project, not a shippable product
- - Still no stock visibility at the till itself
- - Pulls the business deeper into an unstructured channel instead of out of it
Design Approach
I designed SalesSync around a simple contract: every number the owner sees has to trace back to a movement someone recorded once, at the source, across all three surfaces. The admin portal is a dense, calm operational surface for inventory and business managers, stock tables with health indicators, a sales queue built for keyboard-speed triage, and dashboards that answer the owner's three daily questions above the fold. The point of sale is a separate, ruthlessly focused checkout experience for front-of-house staff: search, stock, quantity, confirm, two minutes, touch-first, and tolerant of a dropped connection. The public website introduces the product to new businesses in that same visual language. Data visualization favors plain comparisons, this week against last, movers against dead stock, over anything analytically fancy, because that's how these owners actually reason.
A component library of 70+ elements across the admin portal, point of sale, and marketing website, stock tables, order cards, alert banners, target meters, chart patterns, all built on a shared token system for future white-label deployments.
Key Design Decisions
- 1
Split the product into a dense admin portal and a touch-first point of sale instead of one responsive compromise, each context earned its own design
- 2
The two-minute checkout flow as a hard budget for the point of sale: search, stock check, quantities, confirm. Anything slower loses to the till it replaced
- 3
Stock health rendered as plain-language states, healthy, low, dead, with color and thresholds instead of raw counts, so non-analysts can read the warehouse at a glance
- 4
Shelf-life recorded once at receiving and surfaced as a countdown rather than a date someone has to calculate mentally, so near-expiry stock is impossible to miss
- 5
Sales prediction shown as a simple next-week range beside each product's history, not a black-box number, so owners could sanity-check the forecast against what they already knew
- 6
Offline-tolerant order drafting on the point of sale, syncing once connectivity returns, built around patchy power and network as the normal case, not the exception
- 7
The owner dashboard organized around three standing questions, what's selling, what's stuck, who's performing, instead of a generic KPI grid
Point of Sale · Front-of-House Staff
The checkout surface front-of-house staff use to build orders, check stock, and take payment, tuned for a two-minute transaction under queue pressure. A category-filtered product grid feeds a running order summary, stock counts sit directly on each product card so staff see what's about to run out before they promise it, and the dashboard behind it surfaces the day's amount, order count, and recent activity at a glance.
Admin Portal · Inventory & Business Managers
The operational home for inventory managers, business managers, and executives, dashboards on one shell so each role sees only the numbers that are theirs to act on. Business managers get branch-level sales and inventory analysis, executives get overall revenue and cross-branch comparisons, inventory managers work a full product table with stock levels flagged before they become write-offs, and a full sales history log ties every order back to the branch and rep that closed it.
Marketing Website
The public-facing site that introduces SalesSync to new businesses: a hero that leads with the sales dashboard rather than a stock photo, a features walkthrough of the point-of-sale, e-commerce, and social selling tools, and a pricing and testimonial section built to move a retailer from curious to signed up in one sitting. Full scroll capture below.
Method
Moderated usability testing plus a 3-week field pilot
Participants
12 users across 3 businesses (owners, admin portal staff, point of sale staff)
Task Success
88%
Test Tasks
- 1As point of sale staff, check stock and complete a customer order at the till
- 2As ops, receive a delivery against a purchase order and confirm stock updates
- 3As ops, process the order queue and allocate stock across two orders competing for the same SKU
- 4As an owner, identify the top mover and the worst dead-stock item this month
- 5Set a reorder point and verify the alert triggers correctly
- 6As an inventory manager, log a shelf-life date at stock entry and confirm the near-expiry alert fires on schedule
Key Findings
- 1Reps completed in-shop orders in an average of 1 minute 50 seconds during the field pilot, comfortably inside the two-minute budget that determined adoption
- 2The allocation flow confused ops users whenever two orders competed for the same stock. The interface needed to make that tradeoff explicit instead of hiding it
- 3Owners gravitated to the dead-stock report almost immediately. One pilot business found weeks of freed-up working capital in the very first session
- 4Plain-language stock states tested far better than raw quantity tables built from the exact same underlying data
- 5The shelf-life countdown caught two near-expiry batches in the first pilot week alone, batches that would previously have gone unnoticed until they'd already spoiled
- 6Reps asked for visit-day summaries. Seeing their own totals after a route turned into an unplanned motivation feature
01 · Stock allocation conflicts were invisible
Before
When two orders competed for the same SKU, the system silently allocated on a first-come-first-served basis, and ops users had no way to see or override that tradeoff.
After
I designed an explicit allocation view showing competing orders side by side with priority controls. Ops users started resolving conflicts in seconds, and trust in the queue came back.
02 · Rep order flow stumbled on search
Before
Product search required exact names, but staff actually knew products by informal shorthand, which pushed the checkout flow past its time budget.
After
I added alias support, recent-products shortcuts, and per-customer reorder lists. Average order time dropped below the two-minute threshold in the pilot.
03 · Dashboard buried the owner's real questions
Before
The first dashboard led with a generic KPI grid, and pilot owners just scrolled past it hunting for movers and dead stock.
After
I restructured it around the three standing questions with drill-downs underneath. Owners' time-to-answer in testing dropped from minutes to seconds.
Order Processing Time
Stockout Incidents
Rep Satisfaction
Reporting Speed
Outcome
SalesSync launched with three pilot businesses and expanded from there. Order processing time fell 55% once the structured point of sale replaced the disconnected till-and-transcription pipeline. Stockout incidents dropped 40% on the back of reorder alerts and live multi-warehouse visibility, and shelf-life tracking started catching expiring stock before it turned into wastage. Front-of-house staff, the biggest adoption risk going in, rated the platform 4.6/5. Owner reporting that once took a week of reconciliation now generates 3x faster, on demand, with a sales prediction sitting alongside every product's history.
What We Learned
In operational tools, adoption gets won at the least glamorous surface. The front-of-house checkout flow mattered more than any dashboard, because bad data upstream poisons everything downstream from it.
Plain-language states beat raw data for non-analyst users. Healthy, low, and dead communicated more than a perfectly accurate quantity table ever did, and the same was true of expiry countdowns over raw dates.
A forecast only earns trust if the user can see the history sitting behind it. Showing the prediction next to past performance mattered more than the model's actual accuracy.
Field pilots surface truths lab testing simply can't. Informal product nicknames, patchy connectivity, and shop-floor interruptions reshaped this design more than any moderated session did.
