Case: musician approached 4 days before close — we said no

A guy emailed us at 22:14 on a Tuesday. National talent contest, indie folk act, four full days left on the public-voting phase. He had a small ad budget ready and asked how fast we could start. We read the brief twice, looked at the contest mechanics, and turned him down by lunchtime Wednesday.

This page is about why. It is also about a thing we get wrong in our industry: the assumption that any paying client is a client worth taking.

The situation, briefly

The contest was a regional Polish-language music competition with a multi-round structure. The artist had cleared the jury round and entered the public-voting bracket on a Friday. Voting closed the following Saturday at 23:59 local time. By the time he reached us, 67 hours of the voting window had already elapsed. He was sitting in fourth place out of twelve finalists, roughly 1,400 votes behind third.

His budget was modest but real. He had a solid press kit, a Spotify monthly listener count north of 8,000, and an Instagram following that was engaged but small. Nothing about him screamed "lost cause." If he had reached us seven days earlier, we would have signed gladly.

He didn't. So we didn't.

Why 96 hours is not enough

Here is the unsexy truth about paid acquisition for vote-based campaigns. Meta and Google ads need a learning phase. Audience targeting needs time to stabilize. Creative needs at least one round of feedback from real impressions before you know what is working. None of this is a marketing-blog cliché — it is in the platforms' own documentation. Meta's delivery system explicitly references a learning phase that typically completes after roughly 50 optimization events, and most small campaigns take three to five days to leave it. You can read the current version in Meta's Promotion Guidelines and the related delivery documentation.

So what happens if you launch ads with 96 hours on the clock?

1. Day one: campaign launches, pixel fires, learning phase begins. Spend is inefficient because the algorithm is still guessing.
2. Day two: first creative-performance signal arrives. Bad ads get paused. Good ads get more budget. Cost per click drops.
3. Day three: you finally have stable data. You now have roughly 24 hours of real, optimized spend left.
4. Day four: voting closes mid-day in many time zones. You lose the last evening, which is usually the highest-conversion window.

In a contest where the gap is 1,400 votes and the top three are also pushing their own audiences hard, one optimized day is not enough. We have run the math on this across roughly 30 past engagements. The line where paid traffic starts to reliably move a vote count for a mid-tier candidate sits around six full days of campaign runtime, and that assumes a warm audience and a clean landing experience.

Four days is below the line. It is not a question of effort. It is arithmetic.

The honest version of "we can try"

We could have taken his money. He was willing. He was the one asking. Every agency mind in the room had a version of the pitch ready: "We will do our best, no guarantees, here is your contract." That sentence is technically defensible and morally lazy. It is the polite way of saying, "we will collect your fee and let you discover the result yourself."

That is not what we want to be. We are a four-person operation in Minsk. Our entire reputation runs on whether the next musician, photographer, or pet-grooming finalist who finds us through word of mouth gets honest counsel. One bad story spreads further than ten quiet wins. Read more about how we think about this on our about page.

So we wrote him back at 11:40 the next morning. The email had three parts: a clear no, the reasoning above stated plainly, and an offer.

What we actually did

We refunded the deposit in full the same day. He had paid 40% upfront through our standard intake flow. The transfer went back on the same payment rail, and our refund policy covers this exact scenario — we call it a "scope mismatch refund" internally, and we have done eleven of them in the past two years. It is not a heroic act. It is just operating a service business honestly.

We then offered him a free 30-minute strategy session. Not a sales call. An actual working session focused on what he could do himself in the remaining 96 hours and, separately, how to plan the next contest entry so paid acquisition could actually help. The call happened on a Friday morning. We covered:

• His existing email list, which had 612 unsubscribed-to-newsletter contacts he had never asked to vote.
• Three Polish-language music communities on Facebook where his genre was over-represented and the moderators allowed contest mentions if framed as artist news.
• A simple Instagram Stories sequence he could ship that afternoon, with a clear single call to action and no friction on the vote-redirect page.
• The mechanics of the contest itself: vote frequency, eligibility rules, and what was and was not allowed under the organizer's terms.

That last point matters. We will not help any client work around a contest's terms of service. We will not buy votes, run vote-farm proxies, create sockpuppet accounts, or log into the contest platform on a client's behalf. If your idea of "marketing" requires those things, we are not your agency. There are people who will do that work. They are easy to find, and the contest organizers are getting better at catching them every year. The EU's own framework on unfair commercial practices in Directive 2005/29/EC sets the legal baseline most national contest rules build on, and most public-vote organizers explicitly disqualify entrants whose support is shown to be artificially inflated.

What happened to the musician

He finished sixth. Down two spots from where he was when he contacted us. He emailed back the following Monday — short, generous, said the strategy call had given him a way of thinking about the next round he had not had before. He has not booked us yet. He may never. That is also fine.

The contrarian view we hold, and which costs us business: most contests are not won by spending more in the last week. They are won by the contestants who built a small, warm audience over months and then activated it cleanly in the final stretch. Paid ads to cold audiences in week one of a four-week contest can absolutely move a result. Paid ads to cold audiences with 96 hours left mostly burn budget. We would rather say that out loud than pretend otherwise.

Where this fits in how we work

This case is the reason our intake form asks for the voting close date before it asks for anything else. If a campaign has fewer than six full days remaining, the form routes the inquiry to a human review queue instead of the booking calendar. About 18% of the inquiries we received in the first quarter of 2026 hit that filter. Most of them we declined the same way. A few we took on with a written scope that explicitly limited what we were promising — usually a content-and-organic-only engagement at a reduced fee, with paid acquisition removed from scope because the timeline did not support it.

If you want the longer version of how we set up an engagement, our process page walks through it stage by stage. If you are wondering what fits inside which tier, services and pricing cover it, and the FAQ answers the questions that come up most.

The thing we are protecting is not our schedule. It is the basic claim that hiring us is worth the money. That claim only stays true if we are willing to turn down the engagements where it isn't.

If you are reading this with 96 hours left

Three things, none of them a sales pitch.

First, do not buy votes. The reputational and disqualification risk is real, and the contest organizers' fraud detection has improved noticeably since the major platform tightening around Google's Misrepresentation policy, which has been refined repeatedly through 2024 and 2025. Public-vote organizers borrow heavily from this framing.

Second, work your warm audience. Email list, group chats, Discord, the people who already care. A direct ask from the artist outperforms almost any paid creative in the final 48 hours.

Third, if you want to talk to us anyway — for advice, not a campaign — we will probably take the call. Reach out here. Tell us the contest, the close date, and where you are sitting in the standings. We will give you an honest read in the reply, not on a discovery call.

And if we say no, we mean no. It is not negotiating use. It is the most useful thing we can give you in that moment.