Google Ads offline conversions and Smart Bidding: a 4-week plan for secondary vs primary conversion actions

You've added offline conversions. Now the scary question: what happens to Smart Bidding?

Let me give you the plan I'd follow, because the internet's advice on this is all over the place.

The idea underneath: a conversion action can be primary or secondary. A primary action steers bidding. A secondary action shows up in your reports but doesn't steer anything. So the safe way in is to add the new offline actions as secondary first, watch them, and only promote one when you trust it.

Week one: add the actions as secondary. Don't touch anything else. Check that they're receiving conversions, that the counts look plausible against your CRM, and that the timing makes sense. Most problems show up here: no matches, wrong values, wrong time zone.

Week two: compare. For the same period, put the secondary conversions next to your CRM's real numbers. Are they close? Are the values right? If it's off, fix the plumbing before it steers anything.

Week three: decide which stage should be the goal. Look at volume. If a later stage gets you a few hundred a month, it can be a goal. If it's a handful, it might be too rare, and a middle stage like booked or showed is safer.

Week four: promote one, and change nothing else. Making one primary is the moment bidding starts learning from it. Expect a dip in volume for a couple of weeks while it relearns.

What people typically see after moving a primary goal to a later stage: fewer conversions in the report, on purpose, because a booked call is rarer than a form fill. A higher cost per conversion, also on purpose. A dip in volume while bidding relearns. And, if the stage was well chosen, better leads on the other side.

The mistake to avoid is judging the switch by cost per lead. It will look worse, because you've stopped paying for the cheap junk. Judge it by cost per booked call, cost per sale, or whatever the business actually runs on. And write down cost per lead, cost per booked call and cost per sale before you change anything. Without a "before", nobody can tell whether it worked.

Have a rollback plan before you promote anything. Write down the old primary goal and the date you changed it. If performance falls well beyond the dip you expected, switching back takes a minute, and you'll have a clean before-and-after to learn from.

And don't change budgets or creative in the same window, because you won't know which change caused what. It's boring advice, and it's the difference between a clean experiment and a story you tell yourself afterwards.

A made-up decision to show the reasoning. 300 leads, 90 booked, 50 showed, 20 won. The sale is too thin to steer bidding, so I wouldn't promote it yet. Booked gives 90 examples but includes people who never turn up. Showed gives 50 examples, all of them real prospects. I'd promote "showed", keep "won" as secondary, and revisit in a quarter. The point isn't that showed is always right. It's that the choice comes from counting examples and asking who's in each one.

If you've done this: what happened in weeks one to four? Did you regret it?