If you have ever launched a TikTok campaign, watched the cost per purchase jump around for a few days and felt the urge to change everything, you have met the TikTok Ads learning phase. It is the most misunderstood part of TikTok advertising, and in my experience managing TikTok accounts for brands in Bangladesh and abroad, it is also where most budgets get wasted.
In this guide I will explain how the learning phase actually works, when a campaign exits it, what resets it, and how much you can safely raise your budget without breaking performance. Everything here is based on TikTok’s own documentation plus what I see in live ad accounts every week.
Quick Answer
The TikTok Ads learning phase is the first period after an ad group starts delivering, when TikTok’s system tests different users to find who is most likely to complete your optimization event. According to TikTok, performance typically becomes more stable after about 25 results or 7 days. To get through it, keep tracking accurate, give each ad group enough budget, and avoid major edits until it stabilizes.
- Exit threshold: about 25 results or 7 days (TikTok, June 2026). Older guidance said 50 conversions.
- What resets it: pausing, major edits to budget, bid, targeting or optimization event, and swapping all creatives.
- Budget increase limit: up to 40% per change during learning, up to 30% after learning, with at least 2 days between changes.
- Bid change limit: stay within about 20% of the current bid.
- Practical scaling: 15 to 20% budget steps every 2 to 3 days keep cost per result most stable.
In This Guide
What Is the TikTok Ads Learning Phase?
The learning phase is the first stretch of delivery after you publish a new ad group, or after you make a big change to one. During this time TikTok’s delivery system is still testing who is most likely to complete your optimization event, such as a purchase, a lead or an add to cart. TikTok describes it as the period when the system identifies the best users for your campaign target, and warns that performance may fluctuate while it explores. You can read the definition in TikTok’s official Learning Phase article.
In simple words: in the first days, TikTok is spending part of your budget to learn. Costs are usually higher and less predictable. Once the system has seen enough results, delivery becomes steadier and your cost per result normally settles.
How Long Does the Learning Phase Last?
According to TikTok’s current help article (updated June 2026), volatility typically starts to decline after about 25 campaign results or 7 days from when the campaign enters the learning phase. That is the number to plan around today.
You will still see older blogs and guides mention 50 conversions in 7 days. That was TikTok’s earlier guidance for conversion campaigns using oCPM, and it is still a useful target if your budget allows it. More results simply give the algorithm more confidence. But the official threshold you should plan your budget around now is roughly 25 results within the first week.
Two practical points from real accounts:
- A “result” is your optimization event. If you optimize for Complete Payment, only purchases count. Clicks and add to carts do not help you exit learning for a purchase-optimized ad group.
- If an ad group gets very few results in its first week, it is unlikely to stabilize on its own. Waiting longer rarely fixes a budget or tracking problem.
What Resets the Learning Phase?
TikTok lists three things to avoid while an ad group is learning: pausing campaigns or ad groups, edits that retrigger the learning status, and unreasonable budget settings or creative volumes. In practice, these are the changes I see reset learning most often:
- Large budget jumps, especially doubling a budget overnight.
- Changing the optimization event, for example switching from Add to Cart to Complete Payment.
- Big bid changes on Cost Cap or Bid Cap strategies.
- Changing targeting such as location, age or interests in a way that changes who can see the ads.
- Pausing and restarting an ad group, or turning it off for a few days.
- Replacing all the creatives at once instead of adding new ones next to the old ones.
Every reset sends the ad group back to square one, which means you pay the higher learning-phase costs again. That is why patience in the first week is not laziness, it is cost control.
How to Exit the Learning Phase Faster: Step by Step
Here is the process I follow when I launch a new TikTok campaign for a client. If you are still building your account, start with my TikTok Ads setup guide for Bangladesh first.
Step 1: Make sure tracking is correct before launch
The algorithm can only learn from events it can see. Check that your TikTok Pixel and Events API fire the right event with the right value, and that purchases are not duplicated. If TikTok cannot see a sale, that sale does not count toward the learning phase. Broken tracking is the number one reason I find ad groups “stuck in learning”.
Step 2: Pick an event you can realistically hit
If your store gets two orders a day, optimizing for purchases with a small budget may never reach 25 results in a week. In that case, start one step higher in the funnel (for example Initiate Checkout) to give the system enough signal, then move to purchase optimization once volume grows.
Step 3: Give each ad group enough budget
Your daily budget must be able to buy enough results. TikTok’s budget best practices suggest that for conversion campaigns using Maximum Delivery, the initial daily budget should be at least 20 times your target CPA. Many small advertisers cannot do that, so at minimum make sure the budget can produce around 25 results in 7 days. You will find the math in the example below.
Step 4: Use fewer, stronger ad groups
Splitting a small budget across five ad groups means none of them reaches the threshold. Consolidate. Two well-funded ad groups usually exit learning faster than five starving ones. Keep targeting reasonably broad so the system has room to find buyers.
Step 5: Launch with 3 to 5 creatives and then leave it alone
Give the ad group a few different hooks and angles so it can find a winner. Then do not touch budget, bid or targeting for the first few days. If you need fresh creatives, add them to the existing ad group rather than replacing everything. My TikTok Ads strategy and creatives guide covers what kind of creatives work for Bangladeshi products.
How Much Can You Increase Your Budget?
This is the question I get most often. TikTok’s budget help article gives clear limits:
- While the ad group is in the learning phase: increase the budget by no more than 40% per adjustment.
- After it has exited the learning phase: limit increases to no more than 30% per adjustment.
- Wait at least 2 days between changes so the system can adjust.
- For Cost Cap or Bid Cap, keep bid changes within about 20% of the current bid.
Those are the maximums. In the accounts I manage, I usually scale in smaller steps of 15 to 20% every two or three days. It is slower on paper, but it keeps cost per result stable and avoids the painful “scaled and the CPA doubled” moment. If you need to grow faster than that, duplicate a winning ad group with a higher budget instead of pushing the original too hard.
A Simple Budget Example
Say your target cost per purchase is ৳400. To get about 25 purchases in 7 days you need roughly 3 to 4 purchases per day:
| Item | Calculation | Result |
|---|---|---|
| Results needed | 25 in 7 days | About 3.6 per day |
| Minimum daily budget | 3.6 × ৳400 | About ৳1,450 per day |
| Safer daily budget | Add a buffer for learning costs | ৳1,800 to ৳2,000 per day |
| First increase after exiting learning | ৳2,000 + 15 to 20% | ৳2,300 to ৳2,400 per day |
If ৳1,500 a day is too much for you, optimize for a higher-volume event first, or run one ad group instead of several. The point is to match budget to the number of results the algorithm needs.
Stuck in Learning? What to Check
If an ad group has been running for more than a week without stabilizing, go through this checklist before changing anything else:
- Is the Pixel and Events API receiving your optimization event? Compare TikTok results with real orders in your store.
- Is the daily budget high enough to buy around 25 results a week at your current cost?
- Is your targeting too narrow, for example a single small city with tight interests?
- Is the optimization event too deep for your volume?
- Have you edited the ad group several times this week? Each major edit restarts learning.
- Are your creatives getting clicks but not conversions? Then the problem may be the landing page or the offer, not the learning phase.
If tracking is the issue, my guide on server-side vs client-side tracking explains why browser-only tracking misses conversions, and why sending events from the server helps TikTok see more of your real sales.
TikTok Ads Learning Phase FAQ
What is the TikTok Ads learning phase?
It is the first period of delivery for a new or significantly edited ad group, when TikTok’s system explores different users to find the people most likely to complete your optimization event. Costs and results usually fluctuate during this time.
How long does the TikTok Ads learning phase last?
TikTok says volatility typically starts to decline after about 25 campaign results or 7 days from when the campaign enters the learning phase. Ad groups with very few results may take longer or fail to stabilize.
Does TikTok still need 50 conversions to exit learning?
No. The current TikTok documentation uses about 25 results or 7 days. The older 50-conversion figure is still a good goal if your budget allows, because more results give the algorithm more confidence.
How much can I increase my TikTok Ads budget without resetting learning?
TikTok recommends increasing the budget by no more than 40% per adjustment during the learning phase and no more than 30% after it, waiting at least 2 days between changes. Smaller steps of 15 to 20% are usually the safest.
What resets the TikTok Ads learning phase?
Pausing the ad group, changing the optimization event, large budget or bid changes, changing targeting, and replacing all creatives at once can all retrigger learning.
Should I turn off TikTok ads that look bad in the first 2 days?
Usually not. Early costs are naturally higher. Unless tracking is broken or ads get no clicks, give the ad group at least 3 to 7 days before judging it.
Does adding a new creative reset the learning phase?
Adding one or two new creatives to an existing ad group is much safer than replacing all of them. Large creative swaps act like a fresh start, so add new ads gradually.
Is the TikTok learning phase the same as Facebook Ads?
The idea is similar, but the thresholds and budget rules are different. Follow TikTok’s own guidance instead of copying Meta Ads habits.
Final Thoughts
The TikTok Ads learning phase is not something to fight. Set up clean tracking, give each ad group enough budget to reach around 25 results in a week, keep your hands off for the first few days, and scale in small steps. Do that consistently and your TikTok campaigns will become far more predictable. If you want a deeper overview of the platform, read my complete guide to TikTok Ads in Bangladesh, or see how I use AI to manage campaigns in running TikTok Ads with Claude.
About the Author
Rubel Hasan Rokib is a data-driven digital marketer based in Dhaka, Bangladesh. He manages TikTok, Meta and Google Ads for e-commerce and service brands, and sets up Pixel, Events API and server-side tracking. He is an Upwork Verified Expert with a 100% Job Success score and has worked with more than 200 clients. Learn more on the about page or see his TikTok Ads management service.
Official Resources
- TikTok Business Help Center: About Learning Phase
- TikTok Business Help Center: About budgets in TikTok Ads Manager
- TikTok Business Help Center: Best practices for budget
Need help getting your TikTok Ads out of learning?
I am Rubel Hasan Rokib, a TikTok Ads expert based in Bangladesh. Send me your ad account details on WhatsApp and I will check your tracking, budget and structure, then reply with a clear plan.
3 Comments