SYCGeeks

Vancouver, Canada · Prices in CAD

SYCGeeksSYCGeeks

Support · Apps & APIs · 17 min read

Your first mobile app with this studio

Start at 8–12 screens, one job, both stores, a shared API. Use the screen list this week to see whether you need an app or a site.

AI-generated product device and packaging on a studio table

The decision this week: 8–12 screens, or a site is enough

A first app should do one thing a stranger will pay for or return to: book Thursday, see today’s orders, show the member pass. The useful decision this week is whether that job needs a pocket product of about 8–12 screens, or whether a well-made website with Square booking already finishes it. A marketing wrapper around your site is usually not a reason to go native. SYCGeeks has shipped product work from Vancouver since 2015. The catalog starting point is $28,000 CAD, twelve to twenty weeks, iOS and Android from a shared app, email/password auth, one API the website can use later, and store listing support. Store developer fees and hosting stay separate. You can use the same fence if you never hire us: one job, 8–12 screens, two stores, one API.

Write the first session as a sentence before anyone talks about stacks. “A regular books a 45-minute cut, pays a CAD deposit, and sees it in Square.” That sentence decides navigation, empty states, and whether you need accounts on day one. If two jobs fight — a social feed and a booking — pick one for version 1. The other waits for a quoted add-on after the stores accept you. Native-feeling UI matters more than a novel stack: tabs where people expect tabs, system type where it holds, a thumb that can reach the primary action. If your reference is a consumer super-app, circle the 8–12 screens that are actually your product. The rest is a souvenir.

Builds at this starting price bill in four milestones: 30 percent to start, 30 percent when architecture is approved (the session sentence, the screen list, the API contract), 30 percent when the core works (task + CAD path on TestFlight / internal Android), 10 percent at launch (both stores submitted). Apple and Google review time is not studio time; write days into the calendar and skip scheduling a press day on the same Friday you upload. Discovery is on Pacific time. Book a Square appointment without an account if you want to talk. Write admin@sycgeeks.ca with the job, the deadline, and whether Square, Stripe, or a warehouse already holds the truth. If an existing production app is down, call +1 (672) 200-7924 on a weekday. A first-app brief is not heat. A store rejection the day of a launch is.

Twelve to twenty weeks plus store review does not fit a fantasy date in four weeks. The cheaper path is a dated first version (the 8–12 screens) and a dated second task, not a four-week promise you will miss after flyers are printed. After you approve, the project opens in a dashboard you can read: intake, active, review, paused, launched. You should always know the current step and why a week slipped if it did. A messy brief is usable if it names the job, the stores, the system of record, the languages, and the hard date. An empty brief is not — there is nothing to fence.

Fence the 8–12 screens before you pick a stack

A useful first-version list looks like this: (1) home / next action, (2) sign in or skip, (3) the task list or calendar, (4) the task detail, (5) pay or confirm in CAD, (6) success / receipt, (7) account or member pass, (8) settings and privacy, (9) empty state for a new user, (10) error / offline, (11) a simple operator or history view if the job needs it, (12) a legal/help screen the stores will look for. If you cannot fit the product in that list, you either have two products or you need a website with more room to explain. Live video, wallets as the only rail, or a multi-sided marketplace are a new quote — they are not a thirteenth screen you squeeze in week fourteen.

Onboarding that explains your brand for 90 seconds is how people delete you. Show the task. If they must create an account, say why in one sentence — “so we can show your bookings” — and let them see the calendar first when the product allows it. Magic links and Square customer records beat a password form you will reset every Monday. Empty states are part of the first session: a new user with no orders should see how to place one; a booked user should see the next appointment in Pacific time, with a cancel rule you can defend. Write those states in English and French if both languages are in scope. French strings are longer. Buttons that fit “Book” may not fit “Prendre rendez-vous.” Solve that before store screenshots, not after.

Permissions are asked at the moment of need. Camera when they tap to scan. Notifications when they book and want a reminder. Location when the job is actually nearby, not on first launch. Canadians are tired of apps that demand contacts on install. PIPEDA is about purpose: if you do not need it, skip it, and make the privacy text in the store listing match the privacy page. Offline and bad network are Vancouver facts — parkades, the Seabus, a shop on a concrete block. Cache the last booking, queue a failed submit, show a readable error. “Something went wrong” is not readable. “We could not reach Square. Try again or call the desk.” is readable. That copy is in the project, in both languages if you bought both.

WCAG-minded work still applies on a phone: contrast, labels, dynamic type, and targets a thumb can hit. Test with larger text because operators in their fifties will use the desk build. If a gesture is the only way to cancel, add a button. Accessibility is product quality, not a PDF. A dedicated compliance program is a separate three-to-five-week engagement if you need a punch list beyond the launch bar. Decorative motion can wait. Tablet layouts wait unless the job is a tablet job — a showroom catalog, a kitchen display. Phone first is the default for a first app.

Shared iOS and Android, and one API the site can use later

Ship iOS and Android from one codebase when the product allows it. A first business app — booking, orders, a member pass, a simple catalog — almost always allows it. Two native teams and a second agency for Android is how a 12-week product becomes a 9-month argument about feature parity. You still get platform-native edges: push permission copy, share sheets, the way a back gesture feels. When the product does not allow a shared app — heavy offline maps, unusual Bluetooth hardware, a game loop — that should appear on the CAD quote. Feature flags that leave Android a week behind are how operators lose trust. If Apple Pay is in scope and Google Pay is not ready, say so and keep a card path. Wallets are extras. A card path is the product. GST/HST still applies to what you sell in the app.

Test on a recent iPhone and a mid-range Android, because that is what Canadians actually hold. Skip testing only on the designer’s device. The mobile application includes one API the website can use later: versioned routes, written errors, auth for the clients you actually have. If you already have a site, skip inventing a second catalog. Square, Stripe, or the warehouse stays the system of record for money and stock. The API reads and writes those systems. A fourth truth is how you sell a jacket twice on a Saturday. Webhooks from the processor belong in this design: a paid deposit, a cancelled appointment, a refund should land where operators look, not only in a developer inbox. Retries you can see are the difference between a product and a script.

If you do not have a website yet, still design the API as if a Next.js front end will appear. Operators often need a desk view before they need another consumer surface. Document the contract so a future developer can run it without a phone call. A custom API as its own catalog item is six to twelve weeks from $18,000 CAD with up to three named integrations. Bundled with the app, it sits in the twelve-to-twenty-week path. Ask for partner login in the first brief if you need it — that is auth and a written error contract, priced or cut, not a surprise for finance in week eighteen.

Accounts, notifications, and payments each add a week or more. Say which ones are launch-critical. A booking app that cannot take a CAD deposit is not finished if the brief said it would. A booking app that also wants a social graph, a chat, and a loyalty constellation is a different product. Pick the rail that matches the job: Square when you already take cards in a room, Stripe when subscriptions or a custom checkout live in the app. Push is useful when it is a fact: “Your Thursday 3 p.m. is confirmed.” Marketing push waits for a written decision and a way to turn it off — store reviewers will look for that switch, and so will PIPEDA-minded clients. Quiet hours should respect Pacific time unless you stored another province.

Store listings, review, care, and the twelve-week picture

When the app is handed over, you should have store listing support and a documented API. Screenshots, the subtitle, the first three lines that show before “more,” and the privacy nutrition labels are in the twelve-to-twenty-week path. Apple and Google reject listings that promise a feature the binary does not have, or that use another app’s screenshots. Write what the build does. Canadian English and Canadian French listings are different products if you ship both. Keywords, subtitle length, and screenshot captions change. You need a signer for each language. Machine-translated store copy is how you get rejected or, cheaper to avoid, accepted with a listing that embarrasses you in Quebec. Age ratings, medical claims, and finance language get a tighter review — send the college rule in the brief.

Developer accounts, tax forms, and the legal entity on the listing take calendar time that is not coding. Apple and Google want a Canadian business number, banking, and a person who can click “agree.” Start that in week one. Submitting from a personal account shared with a cousin is the expensive path. If the company name on the listing must match the GST registration, say so. Studio invoices are CAD; the stores have their own fees. App Review is a queue. Submit a build that matches the listing, with a demo account if login is required, and notes that tell the reviewer how to finish the task. “Book the 3 p.m. slot with the demo user, skip payment with code QA-CAD” is a note. “Test the app” is not. Rejections happen — a permission string, a broken privacy link, a payment that does not complete in their region. Fix the defect inside the project. A new feature you thought of while you waited is a quote.

A release cadence you can keep is usually monthly or twice a month, not daily. Leave a checklist: what to smoke-test, who can approve store text, how to read a crash. Care is $450 CAD a month or $4,500 a year: watching crashes, small dependency updates in the reserved hour, a note of what changed. A new screen is an add-on on the same project, about 25 percent less on care, or a block of eight reserved hours if it is bigger than an hour and smaller than a new build. Version numbers should mean something: bump the user-facing version when the task changes. Release notes in human language — “You can now cancel a Thursday booking from the home screen” — beat “Bug fixes and performance improvements” every time. People in Vancouver read those lines. So do reviewers the next time you need a fast review.

The short end of twelve to twenty weeks is for a ruthless product: one task, one processor you already use, English only, no custom hardware. Week one is the session sentence and the API contract. Weeks two to four are the shared shell, auth if it is truly required, and empty states. Weeks five to eight are the task, the CAD payment or Square booking, and the operator view. Weeks nine to twelve are listing copy, both stores, a real charge, and review. The long end is honest when French is in scope, when Stripe subscriptions are the product, when a warehouse feed is dirty, or when you need a desk website in the same breath. Those are not surprises if they are in the brief. Care members pay about 25 percent less on published add-ons; that discount does not make an extra marketplace free. After the stores accept you, watch the first ten real sessions, read the crashes, answer support in the language it arrived in.

CAD payments, GST/HST, and the card path that is the product

If the first session includes money, treat Canadian tax as a screen, not a footnote. Decide in week one: deposit or full charge, which provinces, GST in British Columbia versus HST in Ontario, whether a gift card is in version 1. Write it next to screen 5 on the list. Test a real charge and a refund in CAD before you submit to the stores. A demo card in USD is the expensive path — you will discover the tax line in production. Apple Pay and Google Pay sit beside a card path when the processor can settle them. If one wallet is late, keep the card path; that is how you still launch inside twelve to twenty weeks.

Pick Square when you already take cards in a room and the appointment must appear in the same dashboard. Pick Stripe when subscriptions or a custom checkout live in the app. Receipts must be readable in the language the buyer used. Processor emails are part of the product; if Square cannot send French for that service name, that belongs in the processor decision in week one. GST/HST on a studio invoice (30 percent to start on a $28,000 CAD baseline) is not the same problem as GST/HST on your in-app charge, but both need a named owner before TestFlight.

How to tell the payment is still in scope for 8–12 screens: one product, one tax rule, one refund path, one receipt template. How to tell you have a second product: marketplace splits, store credit balances, usage billing, or a wallet with no card fallback. Those are a new quote. The cheaper path is to ship the card path that matches the chair or the warehouse you already pay for, then date wallets as an add-on after the stores accept you. Care at $450 CAD a month watches the jobs that move money; a new rail is a quote, about 25 percent less on care.

Put a finance signer on the calendar the same week you start Apple and Google legal-entity paperwork. They are often the same person who owns GST registration, and they are rarely the marketing partner. A 20-minute Friday note — “refund path tested in CAD, French receipt still English, wallet waiting” — keeps milestone three (core works, 30 percent) honest. You can run that note without a studio. When you want it on a project record, write admin@sycgeeks.ca or book on Pacific time.

Staff the twelve-to-twenty-week calendar like a product, not a hope

Print this staffing picture for a partner meeting. Weeks 1–3: session sentence, 8–12 screen list, API contract, system of record, languages, developer accounts started, finance named. That is milestone two when architecture is approved (30 percent). Weeks 4–8: shared app shell, empty states, the task, CAD path or Square booking, operator view, tests on a recent iPhone and a mid-range Android. Weeks 9–14: hardening, French strings if in scope, listing copy, screenshots, a real charge and refund, TestFlight / internal testing. Weeks 15–20: store submission, review buffer, cadence checklist, watch the first ten real sessions. Store review is extra calendar, not extra studio days. Skip a press Friday on upload day.

The short end of twelve weeks is for English only, one processor you already use, no custom hardware, one task. The long end is honest when French, subscriptions, a dirty warehouse feed, or a desk website share the breath. Sequence the website (six to ten weeks, about $8,500 CAD, about eight pages) first if operators need a desk before clients need a pocket. Sequence the API (six to twelve weeks, from $18,000 CAD, three integrations) first if partners must connect before either surface. Same project record, named phases. You keep the history. You skip explaining the business to a stranger.

Name a backup signer before week eleven. Content and approvals still move this calendar. A college rule that arrives in week fourteen is a pause or a quote, not a silent extra. Students who touch the build should log hours another person can read: “Square webhook signed, refund path tested in CAD, French listing screenshots exported.” If a week is quiet, ask. A shared note with the current week and the blocker is enough. You can run that without us.

When the stores accept you, watch crashes for ten days, answer support in the language it arrived in, and keep the 90-day defect window for things that broke. A second task in the app is an add-on on the launched project. Fill the screen list this week even if you never hire SYCGeeks — the fence will still tell you whether you are buying an app or a site. If a live app is down tonight, call +1 (672) 200-7924 on a weekday. Otherwise write or book.

App screen list you can fill this week

Copy this list into a note. For each line: in version 1 (yes/no), owner of copy, English string, French string if in scope, empty-state sentence, error sentence. Home / next action. Sign in or skip (why one sentence). Calendar or order list. Task detail. Pay or confirm (CAD, GST/HST behaviour, receipt language). Success. Member pass or account. Settings, notification switch, privacy. Offline / retry. Help / contact (email and weekday phone). Operator or history view. Store listing: subtitle, first three lines, three screenshots, privacy labels, age rating, demo account notes. Systems of record: Square / Stripe / warehouse. Languages on day one. Hard date. Signer and backup. Developer account owner and GST legal name. This list is the cheaper path because a thirteenth “just a feed” becomes visible before you pay 30 percent to start.

How to tell you need a site instead: the first session is “read about us and send a form.” How to tell you need an app: the person already knows you, they repeat a task weekly, they need a pass, a queue, or a payment that should work in a parkade. How to tell you need both: operators need a desk and clients need a pocket — sequence the site in six to ten weeks (about $8,500 CAD, about eight pages) and the app after, on the same project record, so you keep the history. Payments in the app are Canadian payments: CAD, tax behaviour you decided, a receipt the buyer can read in the language they used. Test a real charge and a refund before you submit to the stores. A demo card in USD is not a test. Apple Pay and Google Pay sit beside a card path when the processor can settle them.

Login problems after launch are usually the API or the database, not the customer. Readable errors, a way to write, and a weekday phone for heat belong in the app footer. Wallet login, if you use it on the web, is a different product surface — offer email or the Square identity you already have. Launch support is for things that broke: a webhook that stopped marking Square appointments paid, a French string that overflowed on a small Android, a listing that still shows an old screenshot. New scope — a second marketplace, a third language, a loyalty program — is a published add-on. Folding extra weeks into the original CAD number is how teams lose the meaning of “done.” Ask if you are unsure; the answer should be a sentence.

Start this week: fill the screen list on Monday, book 45 minutes with the signer on Wednesday, start the Apple and Google legal-entity paperwork on Friday even if you have not chosen a studio. Send screenshots of the current process — the paper card, the Square POS, the spreadsheet — to admin@sycgeeks.ca, or book time on Pacific time to walk the first session. Toronto, Montreal, and Halifax clients are welcome on the overlap. You can run the list without us; the 8–12 screen fence will still save you from a super-app you cannot staff. If money is failing in a live app tonight, skip the list and call.

Keep reading