Bid optimisation (Smart Bidding)

This guide explains how to feed high-value signals into advertising agencies' AI algorithms to maximise the effectiveness of your bids.

1. Business Value: Feed AI to perform better

Bidding algorithms (Google Ads, Meta) only perform well if they receive rich data. By sending the margin or customer score rather than just turnover, you teach AI to seek out your most profitable customers (POAS - Profit on Ad Spend).

  • Real Value Signals: Guide AI towards profit rather than gross volume.

  • Fast Learning: Comprehensive data allows algorithms to move out of the learning phase more quickly.

  • Competitive Advantage: Provide data that your competitors do not send (LTV, predictive scores) to win bids on the most promising profiles.

2. Implementation Methodology (Pure Server-Side)

To ensure maximum accuracy without relying on the browser, enrichment is done directly within Platform X.

Step A: Import Segmenting Data

Upload your business data (customer score, VIP status, LTV, margin) to Platform X via a file feed or the Profile API.

Step B: Event enrichment configuration

Use the enrichment engine to intercept your events (e.g. purchase, view_item) and instantly inject the properties from your Profile database.

Step C: Configure the Destination

In your connector (e.g. Meta CAPI), map the enriched property to the conversion value fields or the partner's custom labels.

Step D: Normalisation via Data Cleansing

Use data cleansing to transform your labels (e.g. ‘VIP’) into numerical values or to calculate the final value sent to the advertising agencies.

Step E: Validation via Event Delivery

Verify that your enriched signals are accepted and processed by the advertising networks' APIs.

3. Typical use cases

  1. Smart Bidding on Margin (POAS): Rather than optimising on ROAS, send the actual margin via Server-Side enrichment. The Google Ads algorithm will automatically favour the products that contribute most to your bottom line.

  2. Value-Based Conversion for Lookalike: Send a specific conversion signal only for your high LTV customers. Meta will then be able to model its Lookalike audiences on your best profiles, rather than on the average buyer.

  3. Overbidding on New Customers: Identify non-customers in real time via your Profile database. Send this signal to your bidding engines to increase advertising pressure on the acquisition of new profiles while limiting costs on existing loyal customers.

Need help setting up your bidding signals?

Bid optimisation is the most powerful lever for your ROI. Our experts will help you configure your feeds to maximise the impact of advertising AI. Contact our support team: [email protected]

Last updated