Microsoft Advertising has begun rolling out its largest overhaul of automated bidding models in a decade. The vendor says the rebuilt prediction engine — which adds new user signals and a refined way to value conversions — produces steadier CPA/ROAS performance, faster reaction to target changes and smoother daily budget pacing. For performance marketers, that means a prompt reassessment of tracking, bidding strategy and testing routines.

What changed and why it matters

Automated bidding depends on predicting two things: whether a click will convert, and how much that conversion is worth. Microsoft rebuilt that prediction layer with additional signals and a different conversion-valuation approach. In practice, better predictions should let the system allocate spend more efficiently across inventory and moments that matter.

Microsoft frames the practical outcomes as three advertiser-facing improvements: bids that meet CPA and ROAS targets more consistently without sacrificing conversion volume or value; bids that adjust roughly twice as fast when targets change; and budgets that pace more evenly throughout the day and reach full spend sooner. Those are precisely the operational problems most teams wrestle with when managing daily pacing and target adjustments.

Performance claims — and what to watch for

Microsoft cites internal results: accounts using Maximize Conversion Value or Target ROAS saw, on average, 10% more conversion value per click at a similar ROAS; moving Search campaigns from Target CPA to Target ROAS produced about 8% greater conversion value while keeping ROAS similar; and switching from enhanced CPC (eCPC) to Maximize Conversions with a CPA target delivered roughly 18% more conversions at a comparable cost per action.1–3

These are directionally useful figures, but they are self-reported aggregates. The announcement does not provide sample sizes, vertical mix or controls for seasonality. Treat the numbers as a signal that value-based bidding and improved signals can help, not as a guarantee for every account. You should validate outcomes in your own campaigns under your business conditions.

Action checklist for advertisers

Translate the update into a short, practical plan:

  • Fix conversion tracking first. Automated bidding needs reliable conversion signals. Microsoft recommends Universal Event Tracking (UET) for richer, granular signals, alongside Conversions API and offline imports where needed. If you can, send actual transaction values so value-based bids can optimize to real revenue.
  • Prefer value-based strategies where possible. Target ROAS and Maximize Conversion Value rely on conversion value and tend to benefit most from precise transaction data. If you can map revenue into conversions, test switching from Target CPA to Target ROAS and measure conversion value per click and ROI.
  • Start campaigns with conversion-based bidding at the account level. Microsoft recommends using conversion-based strategies when the account has at least five conversions because account-level history feeds the bidder. If you lack history, begin with eCPC to collect data and switch once you have sufficient conversions.
  • Adjust targets incrementally. Major, sudden changes—new campaigns, switching bid strategies or altering conversion goals—can cause short-term volatility. To reduce disruption, change CPA or ROAS targets in modest steps (Microsoft suggests ~15% increments) and review after a conversion cycle for your business.
  • Avoid hard Max CPC caps. CPC ceilings can conflict with CPA/ROAS objectives and prevent the bidder from pursuing better ROI. Manage outcomes with budgets, bid targets and conversion value rules rather than rigid CPC limits.
  • Evaluate at the right level and with enough data. Assess performance at campaign or portfolio level, not by individual keywords. Microsoft recommends a meaningful sample—ideally ~30 conversions in 30 days—and allowing 1–2 conversion cycles after material changes before drawing conclusions.

How to test this in your account

Run controlled experiments: clone campaigns and apply changes only to test groups while keeping controls unchanged. Track conversion volume and conversion value (not just clicks or CPC), and monitor pacing and budget delivery to validate Microsoft’s steadier-spend claims.

Prioritize value-based experiments where you can supply real transaction values. If you must use proxies for value, document them clearly and expect weaker signal quality. Compare results across multiple conversion cycles and account for seasonality to avoid over-interpreting short-term lifts.

What to watch next

Treat this rollout as an evolving capability. Microsoft says this is the first stage and will continue rolling out automatic improvements through 2027. Watch for three practical signals: detailed release notes that list which new signals were added, independent case studies or third‑party tests that disclose sample methodology, and consistent results in your own controlled experiments. In the short term, focus on clean conversion data, prioritize value-based bids where feasible, and adopt a structured test-and-control approach before making account-wide changes.

For guidance on implementation, consult Microsoft’s help resources on bid strategies: “Let Microsoft Advertising manage your bids with bid strategies.”

1–3 Microsoft internal data: Global, Search and Audience Network, various dates in 2026 (see original vendor announcement for full references).