Custom AppsBuyer's GuideNovember 21, 2026

    How Long Does It Take to Build a Custom App? The Three Finish Lines

    M

    By Mike Evan — Founder, Social Media Strategy HQ•Updated November 2026

    A focused internal web app usually takes four to ten weeks once requirements are written; a customer-facing app with payments and integrations, three to six months; native iPhone and Android apps longer, because device testing and store review add time. Most schedules slip on outside approvals and staff adoption, not on the coding itself.

    Ask a developer how long your app will take and you will get a number of weeks. Ask the same question six months after launch and the honest answer is usually different, and not because the developer was slow. The quoted number measured one thing: how long it takes to make the software work. Two other stretches of calendar sat outside it, and nobody put them on the schedule.

    This guide is about all three. A few related questions are answered elsewhere, so we will not repeat them here. What an app costs at each size, and why the exceptions drive the price, is in our custom business app cost guide. If you are replacing a no-code tool, the migration steps are in when to switch from no-code to a custom app. And the delays that come from the owner's side of any project, such as decisions, content and review rounds, are the same for an app as for a website and are covered in how long it takes to build a website.

    The Ranges, by Who Will Use It

    The single best predictor of an app's timeline is not how many screens it has. It is who logs in. Every step outward, from your own staff to your customers to the public app stores, adds a layer of work that has nothing to do with features.

    An internal tool for your team: four to ten weeks

    A job tracker, a quoting tool, a dispatch board or an inspection checklist used only by your staff. The users are people you can train in a meeting, the devices are ones you control, and if something is confusing on day one, someone walks over and asks. The range assumes the requirements are written down before the clock starts.

    A customer-facing web app: three to six months

    A client portal, an ordering system, a booking tool with accounts. Now each customer must see only their own records, payments have to work every time, and the interface has to make sense to someone who will never be trained. Integrations with accounting, payment or scheduling systems are usually where the extra months go.

    A native iPhone and Android app: longer again

    Two platforms, a range of devices and operating system versions, and a store review for the first release and every update after it. If the reason for going native is offline use, the camera, location or push notifications, those features are also the ones that need the most testing.

    An add-on to a system you already run: often the fastest

    A custom piece that reads from and writes to an existing database, CRM or no-code base can be in use sooner than any of the above, because the data, the logins and the habits already exist. It is the option most worth asking about before committing to something larger.

    One App, Three Finish Lines

    Every custom app crosses three finish lines, and they are not on the same day.

    • It works. The software does what it should, with your real data, for every user who needs it. This is the line a development quote measures.
    • It is allowed. Every outside party that has to say yes has said yes: the app stores, the payment processor, the software vendors whose systems you connect to, and in business-to-business work, your customers' own security reviewers.
    • It is used. The people it was built for have stopped doing the job the old way. The spreadsheet is closed, the paper form is gone, the group text has gone quiet.

    A timeline that only counts the first line is a build estimate, not a project schedule. The gap between the two is where owners lose patience, budgets get re-opened and good apps get labelled as failures they never were.

    Want it done for you?

    Websites, SEO, and AEO — built with Claude Code in days, not months.

    Get a Custom Quote

    Finish Line One: It Works

    The build itself still has a shape, even when it moves quickly. The first week or two goes to pinning down what the app must do, which mostly means listing the cases where the normal process goes sideways. Then the data model, the screens, and the connections to outside systems, each tested as it is built. The last stretch is testing with real users and real records.

    Two things inside this line consistently take longer than owners expect. The first is integrations: each outside system has its own rules, limits and quirks, and each one is a small project rather than a checkbox. The second is your data. A sample of clean records proves very little. The records that break imports are the ones with two phone numbers in one field, the customer entered three times under slightly different names, and the order with a status nobody can explain. Testing with those in the first week is how a schedule stays honest.

    What AI-assisted development shortened, and what it did not

    We build with Claude Code, and the change to this first line is real: screens, data models and integration code that once took weeks can be working in days, and a revision that used to wait for the next sprint can often be made the same afternoon. That is why focused internal tools have moved from a quarter-long project to something measured in weeks. It is also why the other two lines now make up a larger share of every schedule. Faster code does not make an app store reviewer, a payment processor or a busy office manager move any faster.

    Finish Line Two: It Is Allowed

    This is the part of the calendar that belongs to other organizations. None of it is difficult, but almost all of it runs on someone else's queue, and much of it cannot begin until someone at your business fills in a form.

    Developer and app store accounts

    To publish an iPhone app under your company's name, Apple requires an organization account, and enrolling one asks for a D-U-N-S Number for the business. If you do not have one, obtaining it is its own wait. On the Google side, some newer developer accounts must run a closed test with real testers for a set period before an app can go public. Neither is a reason to avoid native apps. Both are reasons to open the accounts in the business's legal name on the first day of the project, not the week before launch.

    Store review, and the resubmission nobody scheduled

    Review itself is usually quick, often a day or two. What costs time is a rejection: a login the reviewer could not test, a feature that needs an explanation, a privacy disclosure that does not match what the app collects. Each one means changes and a new submission. Plan for at least one round, and give the reviewer a working test account from the start.

    Payment and integration approvals

    Payment processors verify the business before money can move, and some software vendors grant access to their systems only after an application or a review of how your app will use the data. Test access often arrives quickly; production access can take longer. Ask each vendor what their approval involves before the build starts, because a finished feature that cannot connect to a live account is not finished.

    Your customer's security questionnaire

    If your app is used by other businesses, expect one of them to send a security questionnaire before their staff are allowed to log in: where data is stored, who can access it, how it is backed up, what happens in a breach. A well-built app answers these easily, but someone has to write the answers, and the customer's reviewer works on their own schedule. Ask your largest client whether they will require one, and start it early.

    Start the slow clocks on day one. Every item in this section can be requested before a line of code exists, and most should be.

    Finish Line Three: It Is Used

    The most common way a finished app fails is quiet. It launches, it works, and three weeks later half the team is still updating the spreadsheet “just to be safe.” Now the business has two systems that disagree, and the new one looks like the problem. Nothing was wrong with the software. Nobody decided when the old way would end.

    The app is finished when the old way is switched off

    Before launch, pick the date the old method retires and tell everyone. Run the two side by side for a short, defined period so the rules nobody wrote down surface while there is still a fallback. Then close the spreadsheet, archive the paper forms and move the group text to read-only. That retirement date is the real end of the project, and it belongs on the schedule from the first meeting.

    Roll out to the people who will complain first

    Start with a small group that does the work every day and will tell you plainly what is wrong, not the managers who will use the app least. Their first week of complaints is the most valuable testing the project gets. Fix what they find, then widen the rollout. For customer-facing apps, the same principle applies: invite a handful of regular customers first, and listen before you announce it to everyone.

    Week by Week: An Eight-Week Internal Tool

    Here is how a focused tool for one team, connected to one outside system, commonly runs when all three finish lines are planned from the start:

    1. Week one: requirements and exceptions written down; access to the outside system requested; a real data sample pulled, messy records included.
    2. Weeks two and three: data model and core screens built; the team's daily user sees a working version by the end of week three.
    3. Weeks four and five: the integration connected and tested against live access; the awkward cases from week one built in.
    4. Week six: a small group uses the app for real work while the old method continues.
    5. Week seven: fixes from that group; everyone else trained; the retirement date confirmed.
    6. Week eight: the old method switched off.

    The building happens mostly in weeks two through five. Everything else is the other two finish lines, which is exactly why they need to be on the plan.

    Can You Pay to Go Faster?

    Partly. Money can shorten the first finish line, up to a point, by adding people or narrowing scope. It does very little for the second, because approvals run on other organizations' queues. It does nothing for the third, because adoption depends on your team's attention, not your budget. The most reliable way to arrive sooner is a smaller first version: one group of users, one job, one or two integrations, launched and used, then extended. A narrow app that people use in eight weeks is worth more than a complete one still waiting on approvals in month five.

    If the first version you have in mind is really several apps, our custom AI apps for small business page explains how we break a larger idea into pieces that can each ship on their own, and AI automation for business covers the cases where connecting the tools you already have gets you there without a new app at all.

    Five Questions to Ask About Any App Timeline

    1. Does this schedule end when the software works, or when the old method is switched off?
    2. Which outside approvals does this app need, and can we request them this week?
    3. When will you test with our messiest records, not our cleanest?
    4. What is the smallest first version that one group of users would actually rely on?
    5. Who on our side has to answer questions, and how quickly do you need those answers?

    A builder who answers the first question with a launch date is only quoting one finish line. If you would rather have a team plan all three, our page on hiring an AI development team explains how those engagements run.

    Get a Schedule With All Three Finish Lines on It

    Tell us who will use the app, which systems it has to connect to and what it is replacing. Social Media Strategy HQ will map the build, the approvals and the rollout on one calendar, then build it for you with Claude Code.

    Start Your App Plan

    Frequently Asked Questions — Custom App Timelines

    How long does it take to build a custom app for a small business?

    It depends on who will use it. A focused internal web app for your own staff typically takes four to ten weeks once the requirements are written down. A customer-facing web app with logins, payments and integrations more often runs three to six months. Native iPhone and Android apps take longer again, because device testing and app store review add their own time. Those figures cover building it; getting outside approvals and getting your team to actually use it are separate stretches of the calendar.

    How long does App Store review take?

    Apple reviews most submissions within a day or two, and Google Play is often similar. The review itself is rarely what delays a launch. What delays it is a rejection that requires changes and a resubmission, a developer account that was not set up in the company's name in time, or a new Google Play account that has to complete a testing period with real testers before it can publish. Start the accounts in week one and plan for at least one round of changes.

    Can a minimum viable product really be built in a few weeks?

    A narrow one can. If the first version does one job for one group of users, works against data you already have, and connects to no more than one or two outside systems, a working version in a few weeks is realistic, especially with AI-assisted development. What cannot be compressed the same way is the time outside the code: account approvals, security questions from your customers, and the weeks it takes for staff to stop using the old method.

    Why does a native mobile app take longer than a web app?

    Three reasons. It has to be built and tested for two platforms and a range of devices and operating system versions. It has to pass each store's review, and every later update goes through review too. And features that justify a native app in the first place, such as offline use, the camera, location or push notifications, are the ones that take the most testing. If none of those features is essential, a web app that works well on a phone usually reaches users months sooner.

    What can we do before the project starts to shorten the timeline?

    Four things save the most time. Request access to every system the app must connect to, in writing, before kickoff. Open the developer and app store accounts in the business's legal name. Pull a real sample of your data, including the messiest records, so it can be tested in week one rather than discovered in week six. And name the one person who can approve changes on the business's behalf, so questions are answered in hours rather than at the next meeting.

    When is a custom app actually finished?

    When the old way of doing the job has been switched off. An app that is live but sits beside the spreadsheet, the paper form or the group text it was meant to replace is not finished; it is a second system. Set a retirement date for the old method before launch, run the two side by side for a short, defined period, and count the project complete on that date, not on launch day.

    M

    Mike Evan

    Founder, Social Media Strategy HQ · Chicago, IL

    Mike Evan is the founder of Social Media Strategy HQ, an AI-first social media agency based in Chicago, Illinois. He works with clients across legal, sports, and business niches to build systematic content and AI-powered marketing infrastructure.