How tROAS actually works
tROAS hands bidding to the platform's model with one instruction: keep the revenue-to-spend ratio near the number I gave you. At auction time the model predicts both the probability of a conversion and its likely value, then bids up when the expected return clears your target and bids down when it does not. It is ROAS used as a steering input rather than a report. The catch is that everything depends on the value data you feed it: if your conversion values are wrong, tROAS optimises confidently toward the wrong goal.
tROAS vs tCPA: value goals vs cost goals
tCPA targets a fixed cost per conversion and treats every conversion as equal. tROAS targets a revenue ratio and cares about how much each conversion is worth. Use tCPA when conversions have roughly uniform value, like a lead or a fixed-price signup. Use tROAS when order values vary widely, like most eCommerce, so the model can chase the 300 dollar carts harder than the 30 dollar ones. Running tROAS without passing real purchase values is the most common way it quietly underperforms.
Why your tROAS target and your real ROAS disagree
A tROAS target is an average the algorithm aims for over time, not a floor it guarantees on any given day. Set 400 percent and you will see days at 250 and days at 600. Set the target too high and the model simply stops bidding, because it can only hit an aggressive ratio by cutting volume until only the surest auctions remain. That is why a target set 50 percent above what the account has ever achieved usually produces a collapse in spend rather than a jump in efficiency.
How to set a tROAS target that holds
The reliable method is to look at the ROAS the campaign already delivers under manual or maximise-conversion-value bidding, then set tROAS at or slightly below that proven number and tighten in steps. Move it 10 to 15 percent at a time and let each change run through a full conversion cycle before the next. iClick never sets a first tROAS from ambition. It sets it from the account's own trailing data, then negotiates the target upward only as the value tracking proves itself trustworthy.
