Most custom builds do not go wrong in one dramatic moment. They drift. A decision waits a week, a file arrives late, a small request gets added without anyone writing it down, and by month three the date has quietly moved. A short, regular check-in catches that early, while it is still cheap to fix and nobody is upset.
This guide is for owners and managers of small and medium businesses whose custom system, web app or mobile app is being built right now. It is general guidance that fits any builder, whether a freelancer, an in-house developer or a software company. It describes no Gotka process, meeting cadence or reporting commitment. How often you meet, how it is held and what the builder reports should be agreed with your own builder, in writing, before the work starts.
Other stages are covered elsewhere and linked rather than repeated. How long a custom system takes explains the phases and typical ranges. Comparing custom software quotes is for before you sign, user acceptance testing is for the end of the build, and bugs versus changes after launch is for after go-live. If your builder has stopped replying altogether, go straight to what to do when a developer goes quiet.
Why a weekly rhythm beats a big update once a month
A custom build is a long series of small choices: what a screen should show, who may approve an order, what a status should be called. Each choice is cheap to change while it is a sketch and expensive to change once other parts are built on top of it. If the owner first sees working software three months in, the feedback arrives as a pile of “this is not quite what I meant”, and every item in the pile is rework.
A short weekly look shortens that loop. Problems are found while they are small, decisions are asked for while they are fresh, and both sides can see whether the plan still matches reality. Weekly is a common starting point rather than a rule. A very small build may need only a written update every fortnight, and a busy final stretch may need a quick call twice a week. Choose a rhythm you will both keep, and then keep it.
The three things to ask for every week
You do not need technical knowledge to run a good check-in. You need to ask for the same three things each time, and a fourth short note on risks.
- A working demo of what changed. Not slides and not “it is going well”, but the real system on a test copy, with the builder showing what is new since last time and you clicking through it yourself. Ten minutes is often enough. If nothing changed, the sheet should say why.
- A short list of the decisions you owe. Each item names the question, the options, who decides, the date it is needed by and what waiting will delay. If you are the person holding things up, you want to know on Monday, not in month three.
- A running list of requested changes, kept apart from the agreed scope. Every new idea goes on the list with who asked, what it is and what it would do to time and cost. Nothing is built from the list until someone has agreed it.
- A note of risks and what each side owes. What could slow the work next week, and what is still outstanding from you: content, data, logins, testers’ time.

A one-page weekly check-in template
Keep the sheet to one page. A page is short enough to be read, filled in and filed, and long enough to hold everything above. Copy the headings into a document in your own company drive so that the history stays with you and not inside the builder’s tools.
| Part of the sheet | What goes in it | Filled in by |
|---|---|---|
| 1. Header | Project name, week number, date, who attended, and the go-live date as it stands today (so any change in the date is visible week by week) | Builder |
| 2. What the builder shows | What changed since last week (shown on the test copy), what was planned but not finished and why, and what is planned for next week | Builder |
| 3. Decisions owed by the owner | The question, the options, who decides, needed-by date, and what waiting will delay | Builder lists, owner answers |
| 4. Change-request list | A number, who asked, what is wanted, the effect on time and cost, and a status: approved, parked or declined | Both |
| 5. Risks and blockers | What could slow the work, who owns it, what would reduce it, and when it will be looked at again | Both |
| 6. What each side owes | Content, data, logins, access and testers’ time still outstanding, each with a date | Both |
| 7. Next check-in | The date and time, and who sends the written summary and by when | Builder |
A simple routine around the sheet keeps it useful:
- Before. Read last week’s sheet. Answer anything you can in advance, and bring the content or data you promised.
- During. Demo first, then decisions, then the change list, then risks. Make decisions on the spot where you can, and write the date beside any you cannot.
- After. A written summary the same day, in the same format, which you read and confirm or correct. That summary is your record if the project later wobbles.
A worked example (made up): week six of a repair-shop job-tracking app
Bright Repair is a made-up appliance repair shop with two service counters. It is having a simple job-tracking web app built. This is how its week-six sheet might look. The names, items and dates are invented for illustration and describe no real project.
| Part | Week six entry |
|---|---|
| Header | Week 6. Attending: owner, builder. Go-live as agreed: end of week 10. No change in the date since week 1. |
| What the builder shows | New job form and a job list searchable by customer phone number, demonstrated on the test copy. Not finished: the technician view, because the status names are still open (see decisions). Next week: finish the technician view once the names are settled. |
| Decisions owed | 1. Final status names (Received, Diagnosing, Waiting for parts, Ready, Collected?). Owner, needed by Friday. Waiting delays the technician view and the reports. 2. Which staff may mark a job Collected. Owner, needed by next check-in. |
| Change requests | CR-04: add a warranty end date to the job form. Asked by owner. Small, adds about one day. Approved. CR-05: print a thank-you receipt. Parked until after the first release. CR-06: send customers a WhatsApp message when a job is ready. Needs its own estimate. Decision next week. |
| Risks and blockers | The old customer list has not yet arrived, so test data is invented. Owner owes it by Monday. Counter staff can only test after 5pm, so the testing window is narrow. |
| What each side owes | Owner: customer list (Monday), status names (Friday), logo file (next Wednesday). Builder: technician view, revised plan if the list is late. |
| Next check-in | Next Thursday. Builder sends the summary the same day. |
Read the sheet as an outsider and you can see what is really holding the project: two answers and a customer list from the owner. The build is not late, but it will be if those three items slip again. That is exactly what a weekly sheet is for. It turns a vague feeling that “things are slow” into three named items with dates.
What slows projects most: late decisions, content and data from your side
It is uncomfortable but common. In many projects the biggest source of delay is not the code. It is waiting for something only the business can supply. Builders can plan around a slow decision for a while, but they cannot invent your price list, your customer data or your approval.
| What arrives late | What it holds up | A habit that prevents it |
|---|---|---|
| Decisions: names, rules, who may approve what | Screens and workflows that depend on the answer, and everything built on top of them | A decisions list with needed-by dates, and a reply time you can really meet, such as two working days |
| Content: text, logos, photos, price lists, templates | Screens that cannot be finished or shown realistically | List the content in week one and give each item an owner and a date |
| Data: customer lists, stock lists, old spreadsheets | Realistic testing, and later the move into the new system. See moving your data into a new system | Send a sample early, tidy the full file in parallel, and agree who cleans what |
| Access: logins and accounts for payment, SMS, email sending, domain, hosting | Connections to outside services, and going live | Open the accounts in your business name early. Account ownership is covered in the stalled-developer guide |
| Feedback on the demo | The next piece of work, which is built on the last | Review within the week and give feedback in one list, not many messages |
| Testers’ time | Finding problems before launch. See user acceptance testing | Book staff time in the plan, not at the last moment |
| New ideas halfway through | The agreed plan and date | Put each on the change list first. Scoping a first app version shows how to sort must, should and later |
The fix is not to apologise for being busy. It is to make each owed item visible, with a name and a date, so that the builder can say “we are waiting on item three” without it sounding like blame. Where a plan assumes the owner will deliver by a certain week, it is fair to ask the builder to write that assumption down.
Warning signs of drift, and what to do about each
A warning sign is a prompt to ask a question, not proof of a problem. Most of these have innocent explanations. What matters is that you notice them in week four and not in month four.
| Sign | What it may mean | What to do this week |
|---|---|---|
| Demos are replaced by status messages such as “going well” | Little working software exists yet, or it is not ready to be seen | Ask for a short demo on a test copy at the next check-in |
| The same item is “almost done” for three weeks | A hidden problem, or no shared meaning of “done” | Ask what remains, in steps, and write down what done means for that item |
| The go-live date moves more than once without a reason | An optimistic estimate, or scope that has grown | Ask for a revised plan that says what changed and why |
| Changes are agreed in conversation with no effect stated | Scope is creeping, one small request at a time | Put every request on the list with its effect on time and cost before work starts |
| The list of decisions you owe keeps growing | You are slow, or questions are being passed to you that the builder could settle | Sort which ones block work, and decide those first |
| Nobody mentions risks any more | There are none, or they are not being discussed | Ask: what could slow us down next week? |
| Payment requests that match nothing you have seen | The payment plan and the progress have come apart | Compare the request with the milestones in your agreement. This guide gives no legal advice, so ask a lawyer before withholding any payment |
| You cannot reach anyone between check-ins | A single point of failure on the builder’s side | Ask for a second contact, and see what to do when a developer goes quiet if replies stop |
Keep change requests apart from the agreed scope
Scope creep rarely arrives as one big demand. It arrives as many small, reasonable requests, each of which sounds too minor to write down. Together they move the date and the cost. The change list on the weekly sheet is the simplest defence, and it is not about saying no. It is about making each yes visible.
- Is it in the agreed scope? If the written scope already covers it, it is part of the work, even if the builder had not yet reached it. If it is something new, it goes on the list.
- Is it a fault or a change? Something that does not work as agreed is a bug to fix. Something that works as agreed but that you would now like different is a change. The maintenance guide explains the one-minute test.
- What does saying yes cost? More time, more money, or something else dropped. Ask the builder to state the effect before you decide, and note it on the list.
- When does it happen? Many requests belong after the first release. Parking a request is a decision, and it should be written down as one.
How easily changes are absorbed also depends on how the work is priced, so it helps to read fixed price or hourly rate alongside your own agreement.
How to run the check-in so it stays short
- Name one decision-maker, and a deputy. If three people each answer differently, the project stalls. The decision-maker holds the final say, and the deputy covers leave.
- Keep to a fixed slot. About thirty minutes is common for a small build. A regular time is easier to protect than a time that changes every week.
- Demo first. Seeing the system keeps the talk concrete and stops it turning into a debate about what was meant.
- Leave faults for the fault list. The check-in is for direction and decisions. A page of small bugs belongs in a separate list that the builder works through.
- Write down decisions with a date. A decision that is only spoken tends to be remembered differently by each side.
- Keep the sheets together. A folder of weekly sheets in your own drive is a record of how the project really went.
- Involve the people who will use it. Bring one or two staff to a demo now and then, as set out in getting your team to use a new system. Their reactions are worth more than guesses.
Agree the routine before the build starts
The easiest time to set up a check-in is before the first line of code, when everyone is still keen. Ask your builder these questions and ask for the answers in writing.
- Where and how will I see work in progress between check-ins, and will it be on a test copy?
- What will the weekly sheet look like, and who writes it?
- Who is my second contact if the main developer is unavailable?
- How are change requests raised, estimated and approved, and how is that reflected in the price?
- What do you need from me, and by which weeks, to keep to the plan?
- How will a changed go-live date be explained to me?
If you are still preparing for the first conversation, what to prepare before you talk to a developer is the place to start. A builder who welcomes these questions is usually one you can work with. One who finds them unreasonable has told you something too.
When drift turns into silence
A check-in routine is a good early-warning system. If sheets stop arriving, demos are missed and replies get slower, the project may be moving from drift to a stall. Do not wait for months. Follow the first-week steps in what to do when a developer goes quiet, which covers writing down what exists, one calm message with a reply-by date and the accounts you should hold in your own name. This guide does not repeat them.
Common mistakes
- Accepting a status message instead of a demo. Words are cheap, and working software on a test copy is the best evidence of progress.
- Treating your own late answers as the builder’s delay. Check the owed list before blaming the date.
- Letting small requests in unrecorded. Ten “tiny” changes are one large one.
- Having several people give the builder different answers. One decision-maker avoids rework.
- Skipping weeks when things look fine. A calm patch is when drift starts, and a five-minute sheet costs little.
- Using the check-in as a bug-report channel. Keep a separate fault list so that decisions get the time.
- Leaving content and data to the end. They are the two things the builder cannot produce for you.
- Keeping no record. Without dated sheets, a disagreement becomes one person’s memory against another’s.
- Waiting until acceptance testing to look closely. By then the cost of change is at its highest.
Where Gotka Technologies fits
Gotka’s App & System Development service builds mobile apps, web apps and business systems, and it is quoted in writing once your needs are scoped. This guide is general advice for any owner with a build under way. It describes no Gotka process: it does not say how Gotka runs projects, how often it meets clients, what it reports or how long a build takes. Ask any builder, Gotka included, to confirm the working rhythm and the reporting in writing before the work starts. If you are planning a new project and want to talk it through, get in touch.
Key terms used in this guide
- Check-in: a short, regular review between the owner and the builder of what was done, what is next and what is blocked.
- Demo: a live showing of the working system, as opposed to a description of it.
- Test copy: a separate version of the system used for trying new work without touching real business data.
- Scope: the written list of what the build will and will not include.
- Change request: a request to add or alter something that is outside the agreed scope.
- Scope creep: the slow growth of a project through many small additions that nobody priced.
- Milestone: an agreed stage of the work with a deliverable you can check and, usually, a payment attached.
- Blocker: anything that stops the next piece of work from starting or finishing, such as a missing decision or file.
- Drift: the gradual slide of date, cost or scope away from the plan without a clear decision to change it.
- Definition of done: a plain statement of what must be true before an item counts as finished.
What should I ask my developer for every week?
Ask for three things: a working demo of what changed, shown on a test copy; a short list of the decisions you owe, each with a needed-by date; and a running list of requested changes kept apart from the agreed scope. Add a brief note of risks and of what each side still owes in content, data and access. Ask for a short written summary the same day.
How long should a weekly check-in take?
For a small or medium build, about thirty minutes is common: a demo first, then decisions, the change list and risks. There is no fixed rule, so agree a rhythm with your builder in writing. A very small project may need only a written update every fortnight, and a busy final stretch may need short calls more often.
What slows a custom software project most?
Very often it is waiting on the client side: late decisions, late content such as text and photos, late data such as customer and stock lists, missing logins, slow feedback and testers who are not free. The builder cannot invent these. Listing each owed item with a name and a date makes the delay visible and easier to fix.
How do I tell a change request from something already in scope?
Check the written scope. If it already covers the item, it is part of the agreed work. If it is new, it is a change request, and it goes on the list with who asked, what it is, and its effect on time and cost. Something that does not work as agreed is a bug, not a change.
What are the warning signs that a build is drifting?
Demos replaced by status messages, the same item almost done for weeks, a go-live date that moves without a reason, changes agreed in conversation with no cost stated, a growing list of decisions waiting on you, and no mention of risks. Each is a prompt to ask a question, not proof of a problem.
What if my builder will not show a working demo?
Ask politely for a short demo on a test copy at the next check-in, and say that you want to see what changed. If none can be shown for several weeks, ask in writing what remains, in steps. If replies also slow down, follow the first-week steps for a stalled developer and consider legal advice about your contract.
Your Software Developer Has Gone Quiet: What to Do in the First Week When a Custom System Build or Support Stalls
Software developer not responding? First-week steps for a stalled custom system: list what exists, send one calm message, check backups, ask for code.
App DevelopmentHow to Compare Two Custom Software Quotes Side by Side
Got two custom software quotes? How to line up scope, inclusions, payment milestones, code ownership and after-launch terms before you compare the price.
App DevelopmentSimple CRM for a Small Business: Spreadsheet, Ready-Made Tool or Custom Build?
Simple CRM for small business in Malaysia: track leads in a spreadsheet, a ready-made tool or a custom build? What to track, signs to switch, and a quick test.


