Your Budget Doesn’t Spend Evenly
GMV Max distributes your daily budget across the day based on predicted conversion probability. The algorithm watches real-time signals — user activity, auction density, predicted purchase intent — and decides hour by hour how aggressively to bid. In theory, this means your spend concentrates where conversions are most likely.
In practice, the algorithm’s predictions aren’t always right. New campaigns lack historical data, so GMV Max enters an exploration phase — spending across a wide range of hours to learn what works. During this phase, some hours burn budget with low conversion rates while the system gathers signal. By the time it identifies your peak window, the daily budget may already be partially depleted.
The result: high-converting hours get starved of budget because low-converting hours consumed it first. Understanding your hourly spend pattern is the first step toward fixing this imbalance.
The Hourly Dimension in GMV Max
TikTok’s Marketing API exposes hourly breakdowns for each campaign. The metrics available per hour include spend, impressions, clicks, conversions, GMV, CPC, CPM, and conversion rate. This gives you a granular view of exactly when your budget is being consumed and what it produces.
One important constraint: hourly data is only available for the previous day (T-1). You cannot see today’s hourly breakdown in real-time through the API. This means hourly analysis is a retrospective tool — you use it to identify patterns over time, not to react in the moment.
AxonRow syncs this data daily and builds a rolling 30-day hourly heatmap for each campaign. Thirty days of hourly data (720 data points per metric) is enough to surface reliable patterns while filtering out day-to-day noise.
Common Patterns
After analysing hourly data across hundreds of TikTok Shop campaigns, several recurring patterns emerge. Your campaigns will likely match one or more of these.
The Evening Spike (UK/US Markets)
Conversions cluster between 7-10 PM local time. This aligns with peak TikTok usage — users scrolling during evening downtime are more likely to watch a full video, engage with a product, and complete a purchase. The intent signal is strong because these users are relaxed and have time to browse.
The problem: if your daily budget is £50 and £30 is spent by 3 PM on low-converting impressions, only £20 remains for the high-converting evening window. The algorithm secured cheap impressions earlier in the day, but those impressions didn’t convert. Now your best hours are underfunded.
The Lunch Dip
The 12-2 PM window often shows high impressions but low conversion rate. Users browse during lunch breaks but don’t complete purchases — they’re distracted, at work, or simply killing time without purchase intent. The algorithm spends here because impressions are cheap and auction competition is lower. But cheap impressions that don’t convert are still wasted spend.
The Late Night Drain
Between 11 PM and 2 AM, you’ll typically see low volume, high CPC, and unpredictable conversion rates. The algorithm sometimes over-indexes on late-night users who show high intent signals (longer watch time, more engagement) but represent a tiny audience pool. The absolute spend is small, but the efficiency is poor — you’re paying premium CPCs for a handful of conversions that could have been acquired cheaper during peak hours.
The Weekend Shift
Weekend patterns differ significantly from weekday. Saturday and Sunday often show earlier peak conversion windows — 2-4 PM rather than 7-10 PM. Users have free time earlier in the day, browse more casually, and make purchase decisions sooner. Sellers running identical budgets and schedules seven days a week miss this shift entirely, applying weekday assumptions to weekend behaviour.
What You Can Actually Do
GMV Max doesn’t offer hourly bid adjustments the way Google Ads does. You cannot tell the algorithm to bid more aggressively at 8 PM or pause spending at 1 PM. But you still have levers:
- Adjust daily budget based on day-of-week patterns. If weekends consistently convert better, increase your weekend budget and reduce weekday spend. GMV Max respects daily budget changes — you can adjust these manually or through the API on a schedule.
- Time your LIVE sessions to match peak hours. For LIVE GMV Max campaigns, your stream schedule IS your ad schedule. If your data shows 7-9 PM is your conversion peak, that’s when you should be live. The algorithm can only promote your stream while you’re broadcasting.
- Refresh creatives during dead zones. If 2-5 PM is consistently low-converting, that’s the ideal window to upload new creatives. The algorithm will test them during these low-stakes hours, gathering performance data without burning your peak-hour budget on unproven assets.
- Monitor budget depletion timing. If your budget consistently runs out before your peak conversion window, you have two options: increase the daily budget so it lasts through peak hours, or switch to Target ROI mode, which paces spend more conservatively across the day rather than front-loading it.
- Compare hourly patterns across campaigns. Different products peak at different times. Fashion and beauty products tend to peak in the evening when users are browsing for inspiration. Household goods and practical items often peak in the morning when users are in problem-solving mode. Running all campaigns on the same budget schedule ignores these differences.
Reading the Hourly Heatmap
AxonRow’s Ads Overview displays a 24-hour heatmap where colour intensity represents conversion rate. Spend bars overlay the heatmap so you can see where budget goes versus where conversions actually happen. The gap between these two layers is your optimization opportunity.
A tight correlation — where spend follows conversions closely — means the algorithm is well-calibrated for your campaign. It has learned your audience’s behaviour and is allocating budget accordingly. This is what a mature, well-performing campaign looks like.
A loose correlation — where spend is distributed relatively flat across the day but conversions spike in specific windows — means budget is being wasted in off-peak hours. The algorithm hasn’t fully learned your pattern, or your daily budget is high enough that it spreads spend broadly rather than concentrating it.
Look for this pattern over a 7-day rolling window minimum. Single-day anomalies (a viral video, a platform outage, a flash sale) will distort the picture. The signal is in the consistency.
When Hourly Data Doesn’t Help
Hourly analysis has clear limitations. For new campaigns with fewer than 7 days of data, there aren’t enough data points for reliable patterns to emerge. You’ll see noise, not signal. For very low budgets under £20 per day, the spend is distributed across too few impressions per hour to draw meaningful conclusions — you might see zero spend in most hours and a single spike where one conversion happened to land.
Highly seasonal products also challenge hourly analysis. If demand shifts week to week (holiday gifting, back-to-school, weather-dependent products), your hourly patterns shift with it. A 30-day rolling window blends multiple demand phases together, obscuring the current reality.
In these cases, focus on daily and weekly trends instead. Hourly granularity becomes useful once you have stable, consistent campaign performance over at least two weeks.
Hourly analysis is one of six reporting dimensions AxonRow tracks for GMV Max campaigns. For the full picture — including creative funnels, product-level Profit ROAS, and optimization mode comparison — see the Ad Analytics playbook chapter.