XHS Ad Bidding Strategy: Manual vs Automatic & When to Use Each
Date Published
Table Of Contents
1. Why Your Xiaohongshu Bidding Strategy Matters More Than Your Budget
2. The Juguang Bidding Framework: What You're Actually Choosing Between
3. Manual Bidding on Xiaohongshu: How It Works and When to Use It
4. Automatic Bidding on Xiaohongshu: How It Works and When to Use It
5. Manual vs Automatic: A Side-by-Side Comparison
6. The Hybrid Approach: Getting the Best of Both
7. Bidding Strategy by Campaign Stage
8. Common Bidding Mistakes International Brands Make
Most international brands entering Xiaohongshu advertising spend hours debating creative direction, targeting parameters, and budget allocation — and then make their bidding decision in under two minutes by accepting the platform default. That's a costly oversight.
Your bidding strategy on XHS determines how Juguang (聚光), Xiaohongshu's official self-serve ad platform, spends your budget in every single auction. Choose the wrong model at the wrong stage, and you'll either overpay for low-intent clicks, burn through budget during the algorithm's learning phase, or miss high-converting users entirely. Choose the right one, and your cost-per-result drops while your reach quality improves — even with the same creative and targeting.
This guide breaks down how manual and automatic bidding work inside Juguang, the specific scenarios where each strategy outperforms the other, and how experienced advertisers sequence between them as campaigns mature. Whether you're launching your first XHS campaign or trying to scale an existing one more efficiently, the decision framework here will sharpen your approach.
Why Your Xiaohongshu Bidding Strategy Matters More Than Your Budget {#why-bidding-matters}
Xiaohongshu operates on an auction-based advertising system. Every time a placement becomes available — whether in the discovery feed or within search results — eligible ads compete in a real-time auction. Your bid isn't the only factor determining whether your ad wins that placement; Juguang also weighs your creative quality score, relevance to the user, and predicted engagement. But your bid sets the ceiling for how aggressively the platform will compete on your behalf.
This has a direct knock-on effect on cost efficiency. A brand bidding manually with well-calibrated amounts can outmaneuver a higher-budget competitor who's on an untrained automatic strategy. Conversely, a mature automatic strategy with deep conversion data can consistently find high-value placements at costs a manual bidder would never discover through trial and error. The bidding model isn't just an administrative setting — it's a strategic lever.
For international brands specifically, the stakes are higher. Entering XHS without prior account history means you're starting with zero platform-side data. The decisions you make in the first few weeks of a campaign set the foundation for how Juguang's algorithm understands your audience, your conversion patterns, and the value of a click to your business.
---
The Juguang Bidding Framework: What You're Actually Choosing Between {#juguang-bidding-framework}
Before comparing manual and automatic strategies, it helps to understand the full menu of bidding models available inside Juguang. They aren't all equivalent — each is designed for a different campaign objective and optimization goal.
CPC (Cost Per Click) is the foundational manual bidding model. You set a maximum bid per click, and Juguang enters auctions up to that ceiling. This model is ideal when you want to drive traffic to a brand profile, a product detail page, or a landing page, and you want direct control over your cost per click from day one.
CPM (Cost Per Mille) charges per thousand impressions rather than per click. This is the go-to model for brand awareness campaigns where reach and visibility are the primary objective, and click-through rate is a secondary concern.
oCPM (Optimized Cost Per Mille) is where the system starts making decisions for you. Rather than simply maximizing impressions, oCPM uses Juguang's machine learning to prioritize placements most likely to deliver your target outcome — whether that's clicks, engagement, or conversions — while still pricing on a CPM basis.
oCPC (Optimized Cost Per Click) works similarly but with a click-based payment structure. The system analyzes real-time conversion probability and competitors' bids, then automatically adjusts your bid on each impression to maximize the chance of a high-quality click within your cost targets.
CPA (Cost Per Action) is the most performance-focused model. You specify a target cost per conversion, and Juguang's algorithm works to hit that target by controlling bids across all auctions. This model requires a well-developed conversion history and pixel integration to function reliably.
No Bid is the most hands-off option — you set a total budget and let the system optimize delivery entirely. This model maximizes budget utilization but offers the least control over cost per outcome.
In practice, the manual versus automatic question maps across these models: CPC and CPM are manual or semi-manual (you set the bid ceiling); oCPM, oCPC, CPA, and No Bid are algorithm-driven, placing them in the automatic category.
---
Manual Bidding on Xiaohongshu: How It Works and When to Use It {#manual-bidding}
With manual CPC or CPM bidding, you set a fixed maximum bid and Juguang uses that figure in every applicable auction. The platform does not adjust your bid based on the likelihood of conversion or the competitive pressure of any given placement — it simply respects your ceiling and either wins or loses the impression.
This sounds limiting, but for international brands entering XHS for the first time, manual bidding is almost always the right starting point. Here's why: automatic bidding models rely on historical account data to optimize effectively. Without conversion history, engagement signals, and audience behavior patterns from your specific account, the algorithm is essentially guessing. Manual bidding removes that dependency and gives you a baseline from which to learn.
Manual bidding is the strongest choice when:
• You're launching a new XHS account or running your first paid campaign with no prior Juguang history
• You're testing multiple creative variants and want controlled, comparable cost data across each
• You're operating in a niche category where platform benchmarks may not reflect your specific audience
• You need strict budget discipline and cannot afford unpredictable spend during a learning phase
• You're running a short campaign (under two weeks) where there isn't enough time for the algorithm to learn
The key discipline with manual bidding is frequent monitoring. Because Juguang won't adjust your bids in response to competitive changes or audience shifts, you need to check performance regularly — ideally every two to three days during an active campaign — and adjust bids when you see consistent over- or under-delivery against your targets.
One practical calibration approach used by experienced XHS advertisers: start your manual CPC bid at roughly 10–20% above the platform's suggested bid range for your category. This ensures your ad enters enough auctions to generate meaningful data without overbidding from the start. From there, you can adjust based on actual CTR and conversion performance.
---
Automatic Bidding on Xiaohongshu: How It Works and When to Use It {#automatic-bidding}
Automatic bidding delegates the moment-to-moment bid decision to Juguang's algorithm. In models like oCPC and oCPM, the system continuously processes signals — your creative engagement rates, audience behavior, competitor activity, time of day, and conversion probability — and adjusts your effective bid on every impression accordingly. In CPA mode, it's targeting a specific outcome cost and managing all bids in service of that goal.
The strength of automatic bidding is its ability to find patterns that no human manager can replicate at scale. An oCPC campaign can identify that users in a specific behavioral segment who engage with certain content types convert at 3x the rate of the broader audience — and reallocate bid weight accordingly, all in real time. This is why mature automatic campaigns frequently outperform manual ones on both cost efficiency and conversion volume.
However, that performance depends on one critical prerequisite: data. Automatic bidding models need historical signals to train on. Without enough conversion events, the algorithm cannot identify what a valuable user looks like for your brand, and it will either overspend on poor placements or under-deliver entirely.
Automatic bidding performs best when:
• Your campaign has accumulated at least 30–50 conversion events, giving the algorithm enough signal to optimize meaningfully
• You're scaling a campaign that has already proven its conversion logic through a manual testing phase
• You're managing a large number of ad groups simultaneously and need the system to handle bid adjustments efficiently
• Your campaign objective is conversion-focused (lead generation, store visits, product purchases) rather than purely awareness-oriented
• You have Juguang pixel tracking and UTM parameters fully implemented, ensuring accurate attribution data flows back to the algorithm
One critical operational note: when you first activate an automatic bidding strategy, Juguang enters a learning phase. During this period — typically 48 hours to two weeks depending on traffic volume — the algorithm is actively calibrating. Performance will often look erratic or suboptimal during this window. Resist the urge to make significant changes to bids, targeting, or creative during the learning phase, as each change resets the calibration process and extends the time to stable performance.
---
Manual vs Automatic: A Side-by-Side Comparison {#manual-vs-automatic}
Understanding the strengths and limitations of each approach in concrete terms helps you make the right call for each campaign you run.
| Factor | Manual Bidding | Automatic Bidding |
|---|---|---|
| Data requirement | None — works from day one | Needs 30–50+ conversions to optimize |
| Cost predictability | High — you set the ceiling | Variable during learning phase |
| Optimization capability | Limited to human adjustment frequency | Real-time, machine learning-driven |
| Best for | New campaigns, A/B testing, short flights | Mature campaigns, scaling, conversion goals |
| Risk profile | Lower risk of surprise spend | Higher initial volatility |
| Management intensity | Requires frequent manual check-ins | Lower once learning phase is complete |
| Response to competition | Static unless manually adjusted | Dynamically adjusts to auction pressure |
The pattern that emerges is consistent: manual bidding is the right tool for the discovery and calibration phase of a campaign, while automatic bidding is the right tool for the growth and scaling phase — once you have data to feed it.
---
The Hybrid Approach: Getting the Best of Both {#hybrid-approach}
Some of the most effective XHS advertisers don't choose between manual and automatic bidding — they use both simultaneously, applying each where its strengths are greatest.
A practical hybrid structure runs manual CPC bidding on your highest-priority, most-controlled placements — such as branded keyword search ads where you want absolute cost certainty — while deploying oCPM or oCPC on broader in-feed discovery campaigns where the algorithm can find audience patterns you haven't anticipated. This structure keeps your most sensitive spend under direct control while allowing the machine learning system to explore and optimize within the awareness layer.
The hybrid approach also makes sense across campaign stages within the same campaign structure. Many seasoned XHS marketers begin a new campaign with manual CPC across all ad groups, collect two to four weeks of data, identify the highest-performing audience segments and creatives, then switch those proven ad groups to oCPC or CPA while keeping newer or experimental ad groups on manual settings. This creates a rolling optimization process that continuously earns better performance without ever flying blind.
For brands working across multiple XHS ad formats — Aurora (in-feed) and Chengfeng (search) simultaneously — the hybrid model is especially valuable. Search ads naturally lend themselves to manual CPC because search intent is explicit and bid-per-keyword logic is straightforward. In-feed discovery ads benefit more from oCPM once you have engagement data, because the algorithm can find audience lookalikes in ways that keyword logic can't replicate.
---
Bidding Strategy by Campaign Stage {#bidding-by-stage}
Thinking about bidding strategy as a sequence rather than a one-time decision is the most reliable way to build compounding performance improvements over time.
Stage 1 — Launch (Weeks 1–2): Start with manual CPC or CPM. Set bids slightly above the platform-suggested range to ensure sufficient auction participation. Focus on generating impressions and clicks rather than optimizing cost. Your goal here is data collection, not efficiency.
Stage 2 — Testing (Weeks 3–4): Continue on manual bidding while running A/B tests across creative variants, audiences, and placements. Identify which combinations generate the lowest cost-per-click and highest engagement rates. Document your baseline benchmarks — these will anchor your automatic bidding targets later.
Stage 3 — Transition (Month 2): Once your best-performing ad groups have accumulated meaningful conversion data, begin transitioning them to oCPC or oCPM. Set your initial automatic bid targets at your established manual benchmarks, then give the algorithm a full learning cycle without intervention. Keep underperforming or experimental ad groups on manual settings.
Stage 4 — Scale (Month 3 and beyond): With automatic bidding now trained on real account data, focus on scaling budget into your highest-performing ad groups. Introduce CPA bidding for conversion-critical campaigns where you have a clear target cost per lead or purchase. Continue using manual CPC for any new creative or audience tests before they graduate to automatic management.
This staged approach mirrors how high-performing brands on other major ad platforms manage their bidding — but it's especially important on XHS, where the platform's unique content environment and the quality-score dynamics of Juguang's auction mean that poorly calibrated early decisions compound over time in ways that are difficult to reverse.
---
Common Bidding Mistakes International Brands Make {#common-mistakes}
Understanding what not to do is as valuable as knowing the right approach. These are the bidding errors that consistently undermine XHS ad performance for brands entering from Western markets.
Defaulting to automatic bidding from day one. It's tempting to set oCPM, choose a budget, and let the algorithm run. But without prior conversion data, the system has no meaningful signals to optimize toward. The result is budget spent on low-quality placements during the training period, with no reliable baseline for comparison.
Making frequent bid changes during the learning phase. Every significant change to your bidding parameters — bid amount, budget, targeting — restarts Juguang's learning cycle. Brands that adjust campaigns daily during the first two weeks of an automatic strategy never exit the learning phase and never see the algorithm's true potential.
Setting manual bids too low to win meaningful auctions. In competitive categories like beauty and skincare, bids set below the platform's suggested minimum often result in near-zero delivery. The campaign appears to be running but is winning almost no impressions, generating misleadingly low cost data without useful scale.
Using the same bidding model for every campaign objective. Juguang's bidding models are designed for specific goals. Using CPM when your objective is conversions, or CPC when you're running a pure awareness play, means you're optimizing for the wrong signal entirely. Aligning your bidding model with your campaign objective is foundational.
Neglecting pixel and attribution setup before switching to CPA. CPA bidding is powerful, but it depends entirely on the accuracy of conversion data flowing back from your landing pages or XHS storefront. Without Juguang pixel implementation and proper UTM tracking, the algorithm is optimizing against phantom data — and your cost-per-action metrics will be unreliable.
---
Key Takeaways {#key-takeaways}
Xiaohongshu bidding strategy is not a one-size-fits-all decision. The right approach depends on where you are in your campaign lifecycle, how much data your Juguang account has accumulated, and what your primary campaign objective actually is.
• Manual CPC/CPM is the right starting point for any new XHS campaign or new creative test — it gives you cost control and generates the baseline data you need before trusting the algorithm.
• Automatic bidding (oCPM, oCPC, CPA) unlocks its full potential only after 30–50 conversion events; introduce it progressively as your best-performing ad groups mature.
• The learning phase is real — when transitioning to automatic, commit to at least 48 hours to two weeks without significant changes to let Juguang calibrate effectively.
• A hybrid structure — manual for new tests and search ads, automatic for scaled in-feed campaigns — is the approach most consistent with how expert XHS advertisers operate.
• Bidding model alignment with campaign objective is non-negotiable: use the right cost model for your goal, not just the default.
For international brands navigating this for the first time, the learning curve is real — but the decision framework is learnable. The brands that consistently outperform on XHS are not those with the biggest budgets; they're the ones who understand the mechanics deeply enough to make the algorithm work in their favor.
Building a Bidding Strategy That Compounds Over Time
The manual versus automatic question on Xiaohongshu is ultimately a question of timing. Manual bidding is how you learn the platform and your audience. Automatic bidding is how you scale what you've learned. Treating them as competing options — rather than sequential tools in the same strategy — is one of the most common and most costly mistakes international brands make when entering XHS advertising.
The brands winning on Xiaohongshu right now aren't necessarily outspending the competition. They're out-strategizing it — building data systematically, transitioning to automation at the right moment, and letting the algorithm compound the gains from well-structured early campaigns. That's a discipline, not a budget line.
If you're planning your first XHS campaign or looking to restructure an existing one for better ROAS, the bidding layer is worth far more attention than most brands give it. Get the resources, frameworks, and category-specific benchmarks you need at AllXHS Free Resources — and explore industry-specific XHS strategies tailored to your vertical.
---
Ready to Build a Higher-Performing XHS Ad Strategy?
AllXHS is the #1 English-language resource hub for international brands marketing on Xiaohongshu. From bidding frameworks and campaign templates to industry-specific playbooks and expert consultation, we give you everything you need to compete effectively on China's fastest-growing social commerce platform.
**Talk to an XHS Marketing Expert →**
Or explore our free Xiaohongshu resources and industry-specific marketing strategies to start building your XHS advantage today.