Custom Bidding in DV360: A Practical Guide for Performance Teams

What Is Custom Bidding in DV360?

Custom bidding is a Display & Video 360 feature that lets you tell the bidding algorithm exactly which outcomes to value — and by how much. Instead of optimising toward a single standard goal like clicks or a default conversion, you define a scoring formula that assigns weighted value to the actions that genuinely move your business.

Standard automated bidding in DV360 is powerful, but it treats every conversion the same. Custom bidding removes that limitation. It gives performance teams the ability to combine multiple signals — different conversion types, on-site engagement, purchase value tiers, or lead quality proxies — into one composite score that the algorithm then optimises against in real time.

For advertisers running sophisticated funnels, this is the difference between chasing cheap conversions and buying the right conversions.

Why Standard Bidding Isn't Always Enough

Most DV360 line items start with an automated strategy: maximise conversions, hit a target CPA, or target a specific ROAS. These work well when your goal is simple and every conversion is worth the same to you.

The problems begin when reality is messier:

• Not all conversions are equal. A newsletter signup and a completed purchase should not carry identical weight.
• You care about downstream value. A conversion that leads to a high-margin product matters more than one that leads to a discounted item.
• You want to optimise toward proxies. Sometimes the true KPI (a qualified sales lead, a subscription retained for 90 days) happens too far downstream to feed the algorithm directly.
• You have multiple objectives. You might want to balance new-customer acquisition against overall volume.

Custom bidding addresses all of these by letting you encode your business logic directly into the bid decision.

How Custom Bidding Works

At its core, custom bidding relies on a scoring script — a set of rules that produces a numeric value for every potential impression opportunity based on its likelihood to drive the outcomes you define. DV360 then bids more aggressively on impressions with higher predicted scores.

The workflow looks like this:

1. Define your goal. Decide what a valuable outcome actually is for this campaign.
2. Create a custom bidding algorithm in the DV360 interface, tied to the relevant Floodlight activities or conversion data.
3. Build the scoring formula using DV360's scripting logic to weight different events.
4. Assign the algorithm to a line item using a maximise-value or target-based strategy.
5. Allow a learning period for the model to gather signal and stabilise.
6. Monitor and iterate as performance data accumulates.

The Role of Floodlight

Custom bidding scripts read from Floodlight activities. Clean, well-structured Floodlight implementation is the foundation of any successful custom bidding setup. If your conversion tracking is incomplete or inconsistent, the algorithm learns from bad signal. Getting measurement right first is non-negotiable — this is where many teams benefit from expert setup support through managed services.

Types of Custom Bidding Approaches

DV360 supports a few distinct ways to build custom bidding logic, each suited to different levels of complexity.

Approach	Best For	Complexity
Weighted Floodlight actions	Combining multiple conversion types into one score	Low
Custom impression value	Optimising toward purchase value or margin tiers	Medium
Custom scripts	Advanced multi-signal logic and proxy goals	High
Data-driven / ML-based models	Predicting deep-funnel outcomes from early signals	High

Many teams begin with weighted Floodlight actions because it delivers meaningful gains with minimal engineering effort, then graduate to scripts as their sophistication grows.

Building an Effective Scoring Formula

A good scoring formula reflects genuine business priorities. Here's how to think about weighting.

Start With Relative Value

Assign each valued action a number that reflects its worth relative to the others. For example:

• Completed purchase: 10
• Add to cart: 3
• Product page view: 1
• Newsletter signup: 2

The absolute numbers matter less than the ratios between them. The algorithm uses these to understand what to chase.

Incorporate Value Where Possible

If you can pass transaction value into Floodlight, use it. Optimising toward revenue rather than conversion count often reveals that the algorithm should behave very differently — pursuing fewer, higher-value users rather than maximising raw conversion volume.

Use Proxies for Deep-Funnel Goals

When your true KPI happens weeks after the click, find an earlier, correlated action to optimise toward. If qualified leads reliably start with a specific high-intent page visit, weight that action so the model can learn from it while it waits for the real outcome.

Keep It Interpretable

Resist the urge to encode dozens of micro-signals immediately. A formula you can explain in a sentence is easier to debug, defend to stakeholders, and improve over time.

Common Pitfalls to Avoid

Custom bidding rewards discipline. Watch out for these frequent mistakes:

• Insufficient conversion volume. Algorithms need enough events to learn. Low-volume line items may struggle to stabilise.
• Changing the formula too often. Every edit resets the learning process. Give changes time before judging them.
• Ignoring the learning period. Early performance is rarely representative. Plan for a ramp before expecting steady results.
• Over-weighting cheap actions. If a low-value micro-conversion carries too much weight, the algorithm will happily flood you with worthless volume.
• Broken or duplicated Floodlight tags. Garbage in, garbage out — measurement hygiene is everything.

When to Use Custom Bidding — and When Not To

Custom bidding is a strong fit when:

• You have multiple conversion types with genuinely different values.
• You can pass reliable value or quality signals into Floodlight.
• You have enough conversion volume for an algorithm to learn.
• Your standard bidding strategy has plateaued and you need finer control.

It may be premature when:

• Your campaign is brand new with no historical conversion data.
• Volume is very thin.
• Your goal is genuinely simple and standard bidding already delivers.

For teams unsure where they sit on that spectrum, a structured account review can clarify whether custom bidding will pay off. Our co-managed services model is designed to help in-house teams build this capability without handing over full control.

Measuring Success

Because custom bidding optimises toward a composite score rather than a single metric, evaluate it against the outcome that inspired it in the first place. Track:

• Weighted conversion value or blended score performance over time.
• Cost per high-value action, not just cost per any action.
• Downstream quality — did the leads or customers acquired actually convert further down the funnel?
• Volume trade-offs — a smart strategy sometimes trades raw conversion count for higher total value.

The key discipline is to judge the strategy by the goal it was built to serve, not by vanity metrics that it was never trying to optimise.

Getting Started the Right Way

Custom bidding is one of DV360's most valuable capabilities, but it rewards preparation. Before you build a single formula, make sure your Floodlight measurement is clean, your conversion definitions are agreed across teams, and you understand the relative business value of each action you'll be optimising toward.

From there, start simple, give the algorithm room to learn, and iterate deliberately. Teams that treat custom bidding as an ongoing practice — rather than a one-time setup — consistently pull ahead of those relying on default strategies. Explore our full range of DV360 services to see how we support advertisers at every stage of platform maturity.

Ready to Optimise Toward What Actually Matters?

Custom bidding can transform DV360 performance when it's built on solid measurement and clear business logic. If you'd like expert help designing, implementing, or troubleshooting a custom bidding strategy, get in touch with our team — we'll help you turn your real objectives into a bidding algorithm that delivers.