If a project is drifting, the usual instinct is to ask for more updates — which makes it worse. Agree the communication contract on day one, use a status format that takes two minutes to write, give feedback in one prioritised round, and treat revisions as something countable rather than an open tap.
Your project is late, or you think it might be, and you do not know because nothing has been said for four days. The instinct is to ask for more frequent updates. It almost always makes things worse: the freelancer spends their week reporting instead of working, you get more information and no more certainty, and the relationship acquires a supervisory tone it will not recover from. What actually works is structure agreed in advance — a stated rhythm, a report format that costs two minutes, feedback delivered once, and a countable limit on revisions. This guide covers each, and what to do when a deadline slips.
Agree the communication contract on day one
Almost all “bad communication” is actually mismatched, unstated expectations. You assume a reply within the hour; they assume within the day. You assume silence means progress; they assume you will ask if you want to know. Neither is wrong, and nobody said anything, so both of you are quietly filing evidence against the other. Fixing this takes one message before the work starts.
- One channel for decisions. Use the project's own messages for anything that matters — scope, approvals, deadlines. Chat elsewhere if you both prefer, but decisions belong where there is a record, for the same reason they would in any other business relationship.
- A response window, not a response time. “Within one working day” is a promise a professional can keep. “Immediately” is a promise nobody keeps, so agreeing to it just means the first missed reply feels like a breach.
- Working hours, both sides. If they work evenings and you work mornings, that is fine as long as it is known. It is only a problem when you are waiting at 10 am for someone who starts at two.
- One scheduled check-in. A weekly written update, or a short call at a fixed time. Fixed beats frequent: a known slot removes the need for anyone to chase.
- What counts as urgent, and how to reach each other when it is. Naming this is what makes it reasonable for everything else to wait.
Send that as one short message at kick-off and ask them to correct anything that will not work for them. A freelancer who pushes back on the response window is not being difficult — they are telling you something true about their capacity, which is far more useful now than in week three.
The status update format that costs two minutes
Asking “any update?” is expensive. It arrives unpredictably, it demands a written reply composed from scratch, and it signals anxiety, which invites reassurance rather than information. Replace it with a fixed format the freelancer fills in once a week. Four lines, always the same four, so it takes them two minutes and takes you thirty seconds to read.
Done this week: the product listing and filter pages are built and on the staging link. In progress: the cart, roughly two-thirds complete. Next: checkout and the confirmation email, expecting to finish by Thursday. Blocked or waiting on you: I still need the final shipping-charge rules and the four product photos for the sale category.
The fourth line is the one that earns the format its place. It surfaces your own dependencies before they become the reason the project is late, and it reframes the update as a shared piece of admin rather than a report to a supervisor. Once you have two or three of these in a row, you can also see drift directly: if the same item has been “in progress” for three weeks, you have a real signal, and you have it without having asked anyone how it is going.
Feedback that can be acted on
Feedback is where clients accidentally create most of the work they later complain about. Three habits account for nearly all of it, and each has a straightforward fix.
| Instead of | Which causes | Say |
|---|---|---|
| “I don't love it, can you try something else?” | A rebuild in a random direction, then another. | “The layout works. The colours feel too corporate for a kids' brand — can we try something warmer, closer to the reference I sent?” |
| Five messages over two days as things occur to you | Rework on rework, and a date that slips for reasons you caused. | One consolidated list, sent once, after you have looked at everything. |
| A flat list of eleven changes | Two hours on your typo and none on the broken form. | The same list in three groups: must fix, would like, and noted for later. |
| “Make it pop”, “more modern”, “cleaner” | Guesswork, because these describe your feeling, not the artefact. | Point at the element and name the property: this heading, smaller; this spacing, tighter; this photo, replaced. |
The pattern underneath all four: describe the problem in the work, not the emotion in you, and send it once. You are allowed to have subjective reactions — taste is legitimate and you are paying for a result you like. But a reaction becomes actionable only when it is attached to a specific element, and it becomes cheap only when it arrives in a single round.
Revisions are a countable thing
“Until I am happy” is not a scope, it is an open tap, and it damages both sides. The freelancer starts absorbing unpaid work and grows quietly resentful; you lose the ability to tell whether you are being reasonable, because there is no number to compare against.
So count them. Two rounds of feedback per deliverable is a common and workable default: you review, they revise, you review again, and that is the milestone. State it before work starts — it is exactly the kind of clause a short scope of work should name, as set out in the contract and scope of work checklist — and then respect your own number. If a third round is genuinely needed because something was missed on their side, most freelancers will absorb it without a word. If the third round is because you changed your mind, pay for it: it is a change request, and treating it as one keeps the relationship honest.
The counting also protects you in an unexpected way. Knowing you have two rounds makes you review properly the first time, which is the single biggest reason projects finish on schedule.
The first missed deadline
One missed date is data, not a verdict. Plenty of excellent freelancers miss a first estimate, because estimating is hard and your project has unknowns in it. What matters is what happens next, and specifically whether you find out from them or from the calendar.
Ask three questions, in a neutral tone, and do not ask for a reason. “Where has it got to, what is the new date, and is anything blocking you on my side?” The reason is rarely useful and asking for one invites either an excuse or a defensive reply; the new date is what you need, and the third question is where you frequently discover that the delay was waiting on something from you. Then mark it: one slip with a fresh, credible date is normal, and if the fresh date holds you have learned nothing bad.
The thing to watch for is the second slip on the same item, or a new date arriving without you having asked. That is a pattern rather than an event, and it is the point to have a different conversation — about whether the scope was wrong, whether they are overcommitted, or whether this should be descoped so something ships. What you should not do is escalate on the first miss, because the cost of treating a normal slip as a breach is that the freelancer stops telling you about problems early, which is the one thing you most need them to keep doing.
Scope changes you actually want
Not every change is scope creep. Sometimes you see the work and realise something better is possible, and refusing to change anything would be the more expensive mistake. The skill is adding it without resetting the project.
Handle it as its own small unit rather than as an amendment to whatever is in flight. Let the current milestone finish, then add the new thing as a separate milestone with its own price, its own date and its own acceptance test — the mechanics are the same as any other milestone, and milestone payments and escrow for clients covers the funding sequence. Bolting a change onto an in-progress deliverable is what turns one clean delivery into two half-finished ones, and it makes the original date meaningless in a way that is genuinely unfair to the person doing the work.
Ask the question explicitly too: “does this change the date?” Most changes do. Hearing the answer before you commit is what separates a decision from a surprise.
Ending well, and what that buys you
The last day of a project is worth ten minutes of deliberate effort, because a freelancer who finished well with you is the single cheapest hire you will ever make again.
Do four things. Run the final acceptance test rather than assuming, and confirm the handover items — files, credentials, access, documentation — are actually in your possession and not merely promised. Release the final payment promptly once they are; a client who pays quickly is remembered, and the delay costs you more in reputation than it earns you in float. Write an honest, specific review. The rating is the part that carries across to their profile, so that is what affects their next hire; the words stay on the project record, and “delivered on time and flagged two problems before I noticed them” is a far better thing for them to be able to point at than a generic five stars. And ask one question: “would you be available for maintenance or the next phase?” If the answer is yes, you have skipped the entire hiring process for your next project.
Then keep the record. Note who you hired, what they did, and what it cost. Next time you have work, you post to the projects board as a known quantity to a known person rather than starting again from twenty proposals — and if you want to see how the whole cycle fits together before your next one, it is laid out on how it works.
The takeaway: managing well is mostly structure agreed in advance, not attention applied later. Set the channel, the response window, the working hours and one fixed check-in on day one. Ask for a four-line update whose last line is what they need from you. Send feedback once, grouped into must-fix, would-like and noted, and describe the element rather than your feeling. Count revisions. Treat a first slip as data and a second on the same item as a pattern. Then end deliberately: test, confirm the handover, release promptly, review specifically, and ask if they are free next time.