Hi! Click one of our member below to chat on WhatsApp
The tidy, left-to-right funnel is gone. People enter from search, WhatsApp, a store, a QR on a flyer, then bounce to mobile, then desktop, then a call. They compare, pause, return, and expect you to remember where they left off. A generic journey cannot handle that reality. A designed system can.
This is your playbook to kill the generic flow and build journeys that feel personal, responsible, and repeatable.
| Symptom | What users feel | Root cause | Fix that works |
|---|---|---|---|
| High drop-off after a beautiful first screen | “Nice, but irrelevant” | Static path ignores intent or history | Intent detection and branch to the right path |
| Repeat users forced through onboarding | “You do not remember me” | No state memory | Persist progress, fast resume, show what changed |
| Confusing errors and loops | “I cannot finish this” | No recovery paths or fallbacks | Design recovery first, then the happy path |
| One message for everyone | “You are shouting, not helping” | Channel blasts over context | Modular content with tone rules and local parity |
| Arabic reads like a translation | “This is not for me” | English-first build | Bilingual planning from day one |
If the interface is pretty but the next step is unclear or irrelevant, the journey failed.
What the user is trying to do. Examples: pay a bill, check eligibility, compare two plans, book a viewing, renew a permit.
What you know right now. Examples: verified identity, location, language preference, payment readiness, previous attempts.
Who gets what next. Examples: show wallet options if KSA, surface Arabic first if last session was Arabic, skip step three if KYC passed.
Headlines, microcopy, offers, and help snippets with approved variants in both languages.
Clear paths for failures. Examples: weak coverage, payment decline, document mismatch, agent handoff.
Events mapped to outcomes, experiments by hypothesis, and a monthly pattern review.
Build the system once. Let every journey use it.
Goal: first success in under two minutes on mobile. Pattern: prefill from device and past sessions, one action per screen, visible progress, “do later” for anything non essential.
Goal: help a decision, not push a promo. Pattern: side by side compare, fees and terms in plain language, top three differences highlighted, switch to Arabic with no layout break.
Goal: zero doubt. Pattern: show total, taxes, fees, arrival or handover date, wallet and card options relevant to the market, one tap reattempt after decline.
Goal: respect and speed. Pattern: name the issue, show the fix, keep data intact, offer human help with context carried forward.
Goal: make staying the easiest choice. Pattern: reminder with what changed, quick confirm, clear benefits, respectful offers, easy opt out.
Voice matrix
| Situation | Tone | Cues | Do not |
|---|---|---|---|
| Announce a benefit | Upbeat and concise | One line promise, then how to use it | Hype with no proof |
| Explain a requirement | Direct and respectful | Define first, example second | Legalese walls |
| Apologize and fix | Own it and act | “We missed this”, action, time to resolution | Passive voice |
| Guide a task | Coach-like | Steps, time estimate, next screen preview | Nested steps and jargon |
Write this once. Store bilingual examples. Use it everywhere the system writes or suggests copy.
is not a translation task. Pair Arabic and Latin typefaces, plan line lengths, and write to both flows from day one.
differ by market. Wallets, cash options, and building conventions matter.
change traffic and response patterns. Schedule notifications and service promises accordingly.
are part of the journey. Explain why you ask, where data goes, and how to fix issues.
Human means culturally precise, not just grammatically correct.
| Area | Metric | Target idea |
|---|---|---|
| Relevance | Time to meaning | Down week over week |
| Progress | First task completion on mobile | Above 80 percent for target segments |
| Efficiency | Steps per task | Fewer steps with equal clarity |
| Recovery | Error to success rate | Above 60 percent recoveries |
| Support | Contacts per 1,000 sessions | Down with stable satisfaction |
| Parity | EN vs AR satisfaction gap | Less than 10 percent difference |
| Learning | Tested hypotheses per month | At least four meaningful tests |
Do not celebrate response time alone. Celebrate time to meaning.
Pick two journeys with measurable pain. Define intent, states, success metrics, and guardrails. Approve the voice matrix and glossary.
Create the modular content library, decision rules, and recovery playbook. Wire analytics to intents and outcomes. Prototype in Arabic and English together.
Ship to a small audience. Run two experiments per journey. Fix friction. Document new patterns in your design system. Roll out widely.
| Do | Do not |
|---|---|
| Design by intent and state | Force everyone through the same five steps |
| Plan bilingual parity from day one | Translate at the end and hope it fits |
| Write recovery first | Treat errors as an edge case |
| Show fees and timelines clearly | Hide terms behind a download |
| Measure time to meaning | Chase clickthrough vanity metrics |
| Document decisions and data sources | Personalize with guesses and no audit trail |
If you want journeys that feel personal, work in both languages, and move the numbers that matter, Spark can help. We design intent-led systems that teams can run and customers trust.
You need clear rules more than new software. Start with your voice matrix, modular content, simple decision rules, and honest analytics. Tools come after clarity.
Fix the highest value drop-off, usually onboarding or checkout. Cut steps, add prefill, write human microcopy, and design a clean recovery path.
Feed models only approved content, enforce guardrails, and keep humans in the loop for tone and risk. Store the approved result as a new example for reuse.
