Projects — Yahiya Ali Khan

Marketing · Real estate · Feb – Sep 2025

FLIVV Meta ad account

Creative Head — Brand & Performance Marketing

₹30Lspend deployed

₹10 Crattributed revenue

33×return

Dailyoptimisation cadence

The problem

High-ticket property sells on trust and timing, not impulse. Leads are expensive and slow, so an account that optimises for cheap form-fills quietly buys a pipeline of people who will never visit a site.

What I owned

Campaign architecture, audience strategy, budget allocation and daily optimisation — plus the creative running inside it and the CRM the leads landed in.

How I ran it

Structure by intent. Cold prospecting separated from lookalikes built on high-intent events, with a retargeting ladder for people who engaged but didn't enquire.
Budget by stage. Allocation followed where the pipeline was thin that week rather than a fixed split.
Creative on a short leash. Because I read the CPL and cut the next hook, losing creative was replaced in days.
Report on money. With the CRM instrumented, decisions moved off CPL alone and onto cost per qualified enquiry and downstream deal value.

I can show you which campaign a closed unit came from. That's the part most media buyers can't do.

Marketing · Technical services · Oct 2025 – present

Cold-start paid social test

Product Designer & Paid Media · Dubai

AED 2Ktest budget

AED 450Kclosed business

225×return on test

0prior paid history

The problem

No paid acquisition history at all. The question wasn't how to scale — it was whether the channel worked for this category, answered for less than the cost of one job.

What I owned

Offer framing, creative, targeting, the landing experience and how enquiries were handled once they arrived.

How I ran it

A real experiment. Tight geo, one narrow service offer, two creative angles — structured so the result would mean something either way.
Proof-of-work creative. Technical-services buyers respond to the job itself, priced clearly, not to lifestyle imagery.
Speed to a human. High-ticket service enquiries die in queues, so the funnel optimised for first response time over raw form volume.
Counted at the invoice. AED 450K is closed revenue tracked through to invoiced work — not pipeline.

Marketing · Real estate · 2025

Property lead-gen funnel

Concept, copy, design and build

4stages owned end to end

1message, all the way through

Figma→ Framer / Webflow

GA4events on every step

The idea

Most funnels break at the seams — the ad promises one thing, the page says another, the follow-up email sounds like a different company. Every seam costs conversion.

What I built

Ad copy, landing pages, lead forms and follow-up sequences as a single message-matched path: the promise in the ad is the headline on the page is the first line of the follow-up.

Principles I held to

One decision per screen. A property enquiry form asking twelve questions is a form nobody finishes.
Form above the fold, always. On mobile especially — the visitor arrived with intent, don't make them hunt.
Fields that filter. The field set was tuned to remove noise before a rep saw it, not to collect everything possible.
Measured, not assumed. GA4 events on each step so the drop-off had a location.

Marketing · Creative production · 2025

In-house creative engine

Creative Head — produced, not briefed

3formats: static, carousel, video

0agency lead time

Daysiteration cycle

2teams directed

Why it matters

Creative is the largest lever in a Meta account and the slowest thing in most orgs. A brief goes out, an agency returns something two weeks later, and by then the winning angle has fatigued.

How I ran it

I concepted, designed and edited the ads myself, then shipped them as structured A/B tests rather than one-off posts — and directed the marketing and design teams producing alongside me.

The person reading the CPL should be the person cutting the next hook. Collapsing that gap is most of the advantage.

Marketing · Measurement · 2025

Tracking & attribution

Pixel, Conversions API, GA4

CAPIserver-side events

Dedupedbrowser + server

GA4funnel events

CPL→CPAreporting

The problem

An algorithm optimises toward whatever you tell it to count. Count form-fills and it will find you the cheapest form-fills in the country — most of them worthless.

What I set up

Pixel plus Conversions API with deduplicated events, mapped so the platform learns from qualified enquiries rather than raw submissions, and GA4 events across the funnel steps.

Note — this is the least glamorous project on this page and probably the highest leverage one. Getting the event definition right is what made the revenue numbers reportable at all.

Marketing · Organic · 2025

Organic, SEO & reels

The unpaid half of the account

SEOon-page & content

Calendarsocial planning

Reelsshort-form editing

Brandconsistent across both

The work

Search and organic social running alongside the paid account — the same positioning, the same offers, so a person who saw an ad and then searched the brand found a consistent story rather than two different companies.

Why it's here

Paid buys attention now; organic compounds. Running both meant retargeting pools stayed fed without paying for every impression twice.

Design · Real estate · 2025

Real-estate CRM

Designed and shipped

Capturefrom ad to record

Routinglead → right agent

Cadencefollow-up sequences

Pipelinestage design

The problem

Ad spend and closed deals lived in different systems, which meant nobody could say which campaign produced revenue. Marketing reported leads; sales reported bookings; the line between them was an argument, not a number.

What I designed

The CRM the leads land in: capture from the ad forms, routing to the right agent, follow-up cadences, and pipeline stages that reflect how a property sale actually progresses rather than a generic SaaS funnel.

Design decisions worth naming

Stages that match reality. Enquiry, qualification, site visit, negotiation, booking — the site visit is the real gate in property, so it earns its own stage.
Routing before beauty. The first screen an agent sees is "who is mine, who is going cold" — not a dashboard.
Source carried all the way. Campaign and creative travel with the record to the closed deal. Without that, attribution is folklore.

Designing the CRM is why I can report revenue instead of leads. It's the single most useful thing I've built.

Design · Regulated software · 2025 – present

U.S. Customs compliance platform

Product Designer · Exact Fit Technical Services

Regulatederror-intolerant domain

Densedata-heavy screens

Workflowclassification & docs

Livein design now

The domain

Customs classification and documentation. Getting a code wrong isn't a bad user experience — it's a fine, a delay, a held shipment. The design constraint is unusual: the interface's main job is to make the wrong entry hard to make.

How I approach it

Progressive disclosure over dense tables, validation at the point of entry rather than on submit, and empty and error states treated as primary screens instead of afterthoughts.

Note — client work in progress; screens are under restriction. Happy to walk through the thinking and the flows live.

Design · Mobile · 2025 – present

Field-services app

Product Designer · Exact Fit Technical Services

Mobileone-handed use

Offlinepoor-signal tolerant

Evidencephoto & checklist capture

Livein design now

Who it's for

Technicians on site — often one-handed, often gloved, often on bad signal, and never in the mood to fight a form. The office app and the field app are different products even when they share a database.

Design constraints

Large targets, minimum required input, work that survives losing connection, and evidence capture that takes seconds because anything slower simply doesn't get done.

Note — client work in progress; screens are under restriction.

Design · E-commerce · May 2024 – Jan 2025

E-commerce ERP

Senior UX Designer · PTR Technologies

ERPstorefront back office

Conversionpaths redesigned

Merchmerchandising tooling

SeniorUX ownership

The work

The systems merchandisers and operators use behind a storefront. Two threads: the conversion path a shopper actually moves through, and the merchandising tools that decide what they're shown in the first place.

Why it shaped me

Designing a conversion path from the inside teaches you that a campaign's ceiling is usually set by the page and the catalogue, not the ad. I stopped blaming creative for numbers the product was causing.

Note — fuller artefacts from this role are available on request.

Design · Commerce PWA · May – Dec 2024

Promode Agro Farms

Web & mobile progressive web app

PWAweb + installable mobile

8 moproject length

Commercecatalogue to checkout

Repeatreorder behaviour

The work

A farm-to-home commerce experience delivered as a progressive web app so one build served both browser and installed mobile use — catalogue, ordering, and the repeat purchase behaviour that fresh-produce businesses actually live on.

The interesting constraint

Groceries are a reorder category, not a discovery one. The design problem is less "how do we impress a first-time visitor" and more "how does someone rebuild last week's basket in under a minute".

Note — screens available on request.

Design · Cloud FinOps · Jan 2021 – Apr 2024

Cloud cost & anomaly dashboards

UX/UI Designer · Synectiks Inc

3+ yrslongest tenure

Spendbreakdown & trends

Anomalydetection surfaces

B2BSaaS product

The work

Three years designing the screens teams use to understand why an infrastructure bill moved — spend breakdowns, trend views, and anomaly analysis. Dense, numeric interfaces where the entire value is making a spike explainable in seconds.

Why it shaped me

Reading a cost anomaly, isolating the driver, and deciding whether to act is the same loop whether the number is an AWS bill or a Meta ad set. I trained on that loop for three years before I ever owned a budget.

I spent three years designing the screen that explains where the money went. Now I'm the one accountable for it.

Design · Infrastructure UI · 2023

Evergreen topology redesign

12–14 weeks

12–14weeks

Topologynetwork / system map

Redesignexisting product

Complexhigh node count

The work

Rebuilding how a complex system topology is drawn, read and navigated — the class of interface that becomes unusable the moment real customer data arrives and a tidy twelve-node diagram turns into four hundred.

The design problem

Legibility at scale: grouping, progressive detail, and giving someone a way to answer "what is connected to this, and what breaks if it fails" without reading the whole map.

Note — screens available on request.

Design · Enterprise internal tool · 2022

ICRISAT procurement tool

5 months

5 moproject length

Internalstaff-facing tool

Approvalsmulti-step chains

Researchinstitute context

The work

Procurement tooling for a global agricultural research institute — purchase requests, supplier records and the approval chains they pass through.

The design problem

Internal tools have captive users, which is exactly why they rot. The work was making a multi-step approval process legible: where a request is, who is holding it, and what unblocks it.

Note — screens available on request.

Design · Non-profit · 2021

ASWS · NGO web & mobile

3 months

3 moproject length

Webpublic site

Mobilecompanion app

Donationgiving flow

The work

Public website and companion mobile application for a non-profit — programmes, outreach and the donation flow.

The design problem

Non-profit giving is the original conversion problem: hold attention long enough to build belief, then get out of the way at the moment someone decides to act. Every extra step between decision and donation costs real money.

Note — earliest project listed here; screens available on request.

Fifteen projects. Two disciplines.

FLIVV Meta ad account

Cold-start paid social test

Property lead-gen funnel

In-house creative engine

Tracking & attribution

Organic, SEO & reels

Real-estate CRM

U.S. Customs compliance platform

Field-services app

E-commerce ERP

Promode Agro Farms

Cloud cost & anomaly dashboards

Evergreen topology redesign

ICRISAT procurement tool

ASWS · NGO web & mobile

And when there's no engineer free?

FLIVV Meta ad account

Cold-start paid social test

Property lead-gen funnel

In-house creative engine

Tracking & attribution

Organic, SEO & reels

Real-estate CRM

U.S. Customs compliance platform

Field-services app

E-commerce ERP

Promode Agro Farms

Cloud cost & anomaly dashboards

Evergreen topology redesign

ICRISAT procurement tool

ASWS · NGO web & mobile