Hi!
I’m planning on investing 150K Robux into advertisements, though I have no idea how to map it out. Should I do 10K a day for 14 days? 150K lifetime a week? If anyone’s experienced on this topic, please help me out!
3 Likes
150K Robux is a great budget, but spreading it evenly (10K for 14 days) or dumping it all at once without testing might not be the most effective way to utilize it.
Here is a strategy that many successful developers use to trigger the Roblox algorithm:
1. A/B Testing (Don’t guess, test!)
Before spending the bulk of your budget, spend small amounts (like 1K - 2K Robux) on 3 or 4 different game icons or ad variations. Check your analytics to see which one gets the highest CTR (Click-Through Rate).
2. The Weekend Push
The Roblox algorithm rewards games with high concurrent players (CCU). A flat 10K every day might just give you a small, flat player base. Instead, take your winning ad from step 1 and concentrate a larger budget (e.g., 30K - 50K) on a Friday or Saturday when player activity is at its peak. This sudden spike in CCU can push your game onto the Discovery page for free organic traffic.
3. Check Your Retention
After the first big weekend push, pause and check your stats. Are players staying? Is your Day 1 Retention good? If average playtime is very low, save the rest of your 150K, fix the early gameplay/tutorial, and then try again.
Good luck with your campaign!
2 Likes
Honestly, I’m pretty bad at making ads, but I’ll give you some of the data I have on a particular game. Do what you will with it.
Ad 1 (4/21-4/25):
- Length: 5 days
- Spend: R$18,525 (~R$3,705/day)
- Peak Impressions (4/26 - 5/14): 889
- Impressions on 10th day (4/30): 66
Ad 2 (5/14-5/21):
- Length: 7 days
- Spend: R$25,080 (~R$3,582/day)
- Peak Impressions (5/22 - 6/15): 23,067
- Impressions on 10th day (5/24): 7,042
(this data is all from the same game, so there’s minimal analytic discrepancies between the two ads)
You should start receiving impressions on the 5th day of ads, meaning that your ad length should be longer than 5 days. Just know that your ad shouldn’t end until a bit after you begin to receive impressions. Aside from that, I don’t have any advice to give. Whether you should go 7 days or 14 days, I don’t know.
2 Likes
I have a question for the weekend push.
Roblox has updated their ad manager (it now uses AI to pick who gets the ad) and it now needs around 24 hours to train for the right audience. Would I just do a lifetime with 50K from Friday evening to Saturday night? Or did I miss something.
Thank you!
1 Like
Yeahh, I was thinking of starting the ad campaign this Monday so it has all the weekdays to train and therefore gives the right audience on a Friday. Thanks!
1 Like
You’re exactly right—Roblox’s new AI-driven ad manager requires a “learning phase” to identify which users are most likely to interact with your specific game.
Feeding a large budget (50K) into a short weekend window is risky because the AI might spend the first 12-18 hours just “exploring” the audience, leaving very little time for the “optimized” delivery phase.
Here is the professional approach to this:
1. Avoid the “Weekend-only” Trap:
If you start the ad on Friday night, the AI will still be in the training phase by Saturday morning. You’ll likely see a lower Click-Through Rate (CTR) and a higher Cost-Per-Click (CPC) during those first 24 hours.
2. The Better Strategy:
Instead of a single “Lifetime” run over the weekend:
- Extend the duration: If possible, try to start the ad on Thursday evening. This gives the AI the required ~24 hours to train before the peak weekend traffic hits on Friday night and Saturday.
- Daily Budget vs. Lifetime: Using a “Daily Budget” often provides more stable delivery than “Lifetime” for a new campaign, as it prevents the AI from overspending too aggressively during the initial exploration phase.
3. The “Cold Start” Reality:
If you must run it only over the weekend, accept that the first 24 hours will be less efficient. Think of that initial spend as the “cost of training” rather than the cost of acquiring players.
In short: You didn’t miss anything, you just correctly identified a flaw in the current AI delivery cycle. If you can’t start early, expect lower returns on Friday night and plan for the bulk of your player acquisition to happen on Saturday and Sunday once the model stabilizes.
1 Like