Google Renamed tCPA and tROAS. Your Mental Model Still Needs Updating.

Google brought back Target CPA and Target ROAS as named bidding strategies in Google Ads. If your first reaction was relief, slow down. A name returning does not mean the underlying machine returned with it. The auction logic that runs beneath those labels has been rewritten at least twice since the original tCPA era. What you are inheriting is a familiar interface wrapped around a system that behaves differently than the one you probably trained your intuitions on.

The Rename Is Not the Story

Google spent roughly 18 months pushing advertisers toward abstracted bidding labels like 'Maximize Conversions with a Target' instead of tCPA. The stated rationale was simplification. The practical effect was that advertisers lost a shared vocabulary for discussing bid constraints with their teams, agencies, and platform reps. The rename restores that vocabulary. It does not restore advertiser control. The model still decides how to interpret your target, when to exceed it, and how aggressively to explore new conversion territory before pulling back. Your tCPA input is a soft guardrail. Probably always was.

What the Rename Actually Signals

Read this as a product positioning move, not an engineering one. Google is consolidating its bidding surface under Performance Max while keeping Search and Shopping campaigns legible to the roughly 73% of ad budgets already running on automated bidding. Legibility matters when those operators are doing budget reviews with CFOs who ask 'what are we optimizing toward.' The tCPA label answers that question. The underlying system answers a different one: how much signal can we extract from this account's conversion history to hit the target often enough that you don't cancel the campaign.

There is a latency problem buried in this. Smart Bidding systems require conversion data to calibrate. New products, seasonal categories, and any campaign with thin historical data will see the model in a learning state longer than most operators budget for. The rename does nothing to accelerate that. If your account has fewer than 30 to 50 conversions per month in a given campaign, the model is essentially guessing. A better label does not change that math.

The Operator's Decision

The question is not whether to use tCPA or tROAS. Most mid-size and enterprise accounts are already locked into automated bidding by default. The question is how deliberately you are setting the inputs that constrain the model. A tROAS target that is calibrated against your actual margin curve behaves differently from one set to match a competitor's reported benchmark. A tCPA set against blended conversion value behaves differently from one built on a single high-intent action. The rename gives you back a shared language. Use it to run tighter internal alignment on what those targets actually represent, not just what Google's interface suggests.

Three Implementation Steps Worth Running Now

First, audit which campaigns have target values set and when those values were last updated against real margin data. In many accounts, tROAS targets were set during onboarding and have not moved since. Second, map your conversion actions to the targets that reference them. If a tCPA campaign is optimizing toward a micro-conversion like 'add to cart' rather than 'purchase,' the model is hitting its target while your revenue misses. Third, document your bid target rationale somewhere outside the platform. Vendor lock-in is partly cognitive. When your account gets reassigned to a new internal owner or agency team, the reasoning behind your targets should survive that transition.

Three questions to pressure-test your bidding setup

Does your current tROAS or tCPA target reflect a margin threshold, or did someone anchor it to a number the platform suggested at campaign creation? If Google's model has a 30-day learning window in your category, how many of your active campaigns are operating inside that window right now, and what decisions are you making based on their in-flight data? When your target is exceeded for three consecutive days, do you have a documented protocol for whether to raise the target, add a budget cap, or hold steady and let the model correct? One honest uncertainty: it is not fully clear whether the naming restoration reflects internal Google feedback that abstracted labels hurt campaign performance, or whether this is purely a sales and retention response to advertiser complaints. That distinction matters for inferring what other UI changes are coming. If Google releases data showing tCPA-labeled campaigns outperformed abstracted equivalents on retention or spend growth, that would change the interpretation substantially.