What changed
Microsoft Advertising used its Activate 2026 event (May 19, 2026) to ship a concentrated set of automation and reporting features that, taken together, represent the most aggressive push Microsoft has made to match Google's AI-driven ad stack.
The headline feature is AI Max for Search, currently in closed pilot. It bundles three things: expanded query matching (closer to broad match in behavior), real-time creative personalization based on the user and context, and built-in experimentation tools so you can test it without torching your existing structure. It's Microsoft's direct response to Google's AI Max for Search.
On measurement, Custom Columns now support conversion metrics, meaning you can construct metrics like lifetime value (LTV, the projected long-run revenue from a customer) and average order value (AOV) directly in the reporting UI. Previously, getting to profit-based optimization required exporting data and doing the math outside the platform.
Performance Max picked up three meaningful controls: New Customer Acquisition goals oriented around incrementality (conversion lift attributable to the ad, not just correlation), expanded exclusions so you can fence off audiences or placements, and publisher-level transparency showing where your ads actually serve. Microsoft says PMax with these goals delivered roughly an 8% lift in incremental conversions. That number is self-reported, so treat it as a directional signal rather than a guaranteed outcome.
Copilot Root Cause Analysis automates the diagnosis of performance drops. It reads account signals, surfaces likely causes, and flags conversion-tracking breakage. Product Explorer landed in Merchant Center in June 2026, giving a searchable catalog view that shows which products are serving, which have feed issues, and which are actually driving results.
Rounding out the announcements: Bid Strategy Reporting and Cross-Account Portfolio Bidding for larger accounts, Impression-Based Remarketing (targeting users based on ad exposure, not just clicks), Content Targeting on specific Microsoft properties like MSN and Outlook, and a simplified LinkedIn Targeting interface.
Who it affects
If you're spending under $1k/month on Microsoft, most of this is future-state. The features that matter now are Custom Columns (available to everyone), Product Explorer (Merchant Center users), and PMax controls (if you're already running PMax).
If you're spending $5k/month or more, or if you manage multi-brand accounts, the full stack is relevant today. AI Max pilot access will roll out selectively. Cross-Account Portfolio Bidding directly addresses the structural gap that forced large accounts to manage bids campaign by campaign. LinkedIn Targeting improvements matter if B2B audiences are part of your Microsoft strategy.
Why it matters
Microsoft has historically been the "set it and forget it" platform: solid CPCs in some verticals, but a reporting and automation layer that was years behind Google. That gap is closing quickly.
The Custom Columns expansion is underrated. The ability to build LTV and AOV metrics natively changes the optimization target. Right now, most Microsoft accounts optimize to cost-per-conversion because that's what the platform surfaces easily. If you can surface revenue-per-conversion or LTV directly in the bidding interface, you can push smart bidding toward profit rather than volume. That's the same shift that took Google accounts 18 months to get right when tROAS became mainstream.
AI Max for Search is the riskier one. Looser query matching plus real-time creative changes means your search campaigns will serve on queries and show copy you didn't explicitly write. The built-in experimentation tools are there precisely because Microsoft knows advertisers won't accept that blindly. The right move is to run it as a controlled experiment, not an account-wide rollout.
PMax's new controls matter because the previous version of Microsoft PMax was genuinely hard to trust: limited exclusions, opaque placement data, no new-customer signal. Expanded exclusions and publisher transparency make it closer to a manageable channel. The incrementality-focused New Customer Acquisition goal is conceptually sound, but the 8% lift figure comes from Microsoft's own reporting and should be validated against your own baseline before you restructure campaigns around it.
The play
-
Build LTV and AOV Custom Columns today. Pull your transaction data, calculate your average LTV and AOV by campaign or product line, and set up the columns. Use this to recalibrate any tROAS targets that were set against raw conversion volume. Watch blended CPA and revenue per conversion over 30 days.
-
Request AI Max pilot access and set up a 50/50 experiment. Don't opt your whole account in. Run AI Max against your current search structure at equal budget, let it run to statistical confidence (aim for 95%), and gate the rollout decision on CPA and conversion quality, not just click volume.
-
Turn on PMax New Customer Acquisition goals on one growth campaign. Set a new-customer revenue target, let it run for 30 days, and measure new-customer share of revenue against your prior 30-day baseline. If it doesn't clear the ~8% incremental lift bar Microsoft cited, that tells you something about your audience saturation on the platform.
-
Audit your Merchant Center feed with Product Explorer. Find the products with zero impressions or active feed issues, prioritize the top 10 by potential revenue, fix them, and track impression share recovery over two weeks. Feed quality is still the highest-leverage lever for Shopping performance, and now you have a native tool to find the gaps.
-
If you run a large account, evaluate Cross-Account Portfolio Bidding. The structural benefit is shared learning across campaigns, which matters most when individual campaigns have thin conversion data. The risk is that a poorly performing campaign can drag shared bid strategy performance. Audit conversion volume by campaign before consolidating.
Watch-outs
AI Max's looser query matching will likely surface irrelevant traffic if you don't have tight negative keyword lists in place before you pilot it. Clean your negatives first. Creative personalization also means your brand guidelines need to be locked in via the Brand Controls feature, or you'll see off-message copy in the wild.
Custom Columns are only as good as your underlying conversion data. If your conversion tracking is misconfigured (something Copilot Root Cause Analysis can now flag), building LTV columns on top of bad data just makes the bad data look more official. Fix tracking before you optimize to derived metrics.
The Microsoft-reported 8% incremental conversion lift for PMax is a platform-level average. Branded, high-intent, or retargeting-heavy accounts will likely see lower lift because those conversions would have happened anyway. The goal is built for prospecting, so if PMax is currently your retargeting workhorse, adding this goal without restructuring will muddy your results.
The WhyItWon angle
AI Max, expanded PMax, and real-time creative personalization all increase the number of variables Microsoft controls on your behalf. That's good for accounts that were under-optimized, and genuinely risky for accounts where creative consistency and query intent are the competitive edge. The problem is you won't know which category you're in until you've spent money finding out. WhyItWon reads your existing ads, your rivals' creative, and what's resonating with your customers to score what's likely to win before you commit budget to a new campaign type or a platform's AI making creative decisions for you. When the platform is doing more, knowing what your creative fundamentals are going in matters more, not less.
Industry