A lever pointed at profit, not just conversions
For years we've been optimizing for ROAS. It was a fine proxy while it worked, but ROAS is a revenue metric wearing a profitability costume. It treats a low-margin hero SKU and a fat-margin slow mover as the same thing, because all the algorithm sees is a conversion value. Google's new Product Value Optimization (PVO) beta is the most honest admission yet that this is a problem: a way to point automated bidding at the products that actually pay your bills, rather than just asking the machine to generate more top-line value.
Google quietly shipped the beta in June 2026, and almost nobody noticed until PPC consultant Adriaan Dekker surfaced it on LinkedIn. That's typical. The genuinely useful features in Google Ads rarely arrive with a keynote. They land in a help-doc changelog and wait for someone to read the footnotes.
Here's the plain-English version: PVO lets advertisers configure value adjustments that guide automated bidding toward the products that matter for business goals. Not "products that convert." Products that matter. That gap is where the entire feature lives.
How the value adjustments actually work
The mechanic is simpler than the name suggests. You apply a percentage adjustment at three levels: individual product, brand, or category. The adjustment tells the bid algorithm to behave as if a product is worth more or less than the raw conversion value it sees.
Google's own example makes the intent obvious. Label a product "High-margin" and weight it +32%. A "Standard" product gets +12%. A "Low-margin" item gets -5%. The algorithm then pushes spend toward the high-margin bucket because, in its internal math, those conversions are now worth more.
The numbers come with guardrails, and they're worth knowing before you try to get clever:
- The floor for any adjusted value is 0.1.
- You can increase a product's effective value by up to +1000%.
- You can decrease it by up to -90%.
So the ceiling is high. A product Google scores at 1.0 can be pushed to 11.0 in theory. In practice you rarely need anything that blunt. Most retailers want a nudge, not a catapult.
I'd treat +32% / +12% / -5% as a sensible starting scale rather than gospel. The point isn't to copy Google's demo numbers. The point is that you finally have a dial for relative importance. And the same discipline you'd apply when validating your smart bidding targets applies here: pick adjustment values from your actual margin data, not from gut feel, and give the algorithm time to learn before judging the shift.
Where you set it, and which bidding strategies qualify
You configure the adjustments either at the campaign level (Goal Settings) or at the conversion level (Conversion Settings of a value-based conversion action). That choice matters: goal-level settings apply broadly, conversion-level settings bind to the specific action being optimized.
PVO is available for the value-based bidding strategies:
- Target ROAS
- Maximize Conversion Value
- Maximize Conversion Value with a Target ROAS
It is not available for Maximize Conversions, Maximize Conversion Value with Max Value (when no ROAS target is set), or Maximize Conversions with a Target CPA. That restriction tells you everything about how Google thinks about this: PVO only makes sense when the algorithm is already chasing a value target. You can't steer value if value isn't the objective in the first place.
Your reporting stays clean
The part I appreciate most: these adjustments do not change the actual conversion values reported in your accounts. This is a steering signal on the bidding side, not a data mutation on the reporting side. You are not cooking the books. You are telling the model where you think the real money is, and your reports keep telling you the truth about what actually happened.
That separation is what makes PVO trustworthy enough to actually deploy. Plenty of "controls" in automated bidding quietly distort the numbers you use to judge performance. This one doesn't.
Performance Max needs its own setup
For Shopping campaigns, applying PVO is straightforward — you set the adjustments and you're done. Performance Max requires one extra step. You create a label, add eligible assets to it, then attach the value adjustment to that label group. The same product, brand, or category logic applies; PMax just asks you to do a little more bookkeeping before the adjustments take effect.
It's friction, not a blocker. If you've ever wrangled PMax asset groups — and you've been following Google's parallel experiments with channel prioritization controls for Performance Max — one more label is a minor tax.
The practical playbook
Google frames the obvious use cases, and they map cleanly to real retail problems rather than hypotheticals:
- Push higher-margin products and higher AOV items.
- Surface products with stronger long-term profitability, or higher customer lifetime value.
- Boost bestsellers when you expect a surge in demand.
- Move overstock and clearance inventory while it still has value.
- Give seasonal promotions extra weight.
Notice what those have in common. They are finance-team concerns that we used to solve by reorganizing campaign structure — carving out separate budget buckets, building dedicated SKUs campaigns, fragmenting accounts so the algorithm would pay attention. PVO gets you closer to the same outcome without tearing up the structure.
This is finance meeting performance marketing
Step back from the feature mechanics and the bigger story shows. PVO is the first time Google has let a profitability argument reach directly into the bidding logic. A margin threshold, a cash-flow need, a sell-through target — these can now drive the algorithm instead of being reverse-engineered through budget hacks. Overstock? Push it. Q4 cash crunch? Weight the high-LTV items.
For years we optimized for ROAS because it was the closest number the machine understood. Product Value Optimization is Google admitting that ROAS was always a stand-in for profit, and finally handing us a way to optimize the real thing. It's the natural next step for anyone already working to move PPC reporting beyond search terms toward data-driven profitability metrics: once your organization agrees on what a truly profitable conversion looks like, PVO is the dial that feeds that definition straight into the auction. The retailers who treat their product feed and their margin data as living optimization inputs — not a one-time setup — are the ones this was built for. Turn it on and use it deliberately.
For the underlying announcement and configuration details, see the original Search Engine Land coverage.
