new app campaign: max clicks or conversions?

jamesaso

So I've got 300+ install conversions across multiple app campaigns-UAC, branded search, the lot. They all flatlined, so I spun up a fresh search campaign. Conventional wisdom says you need clicks first if you've got zero history, but my account's stacked with conversion data. still, I ran it on Max Clicks for two weeks... got exactly one install. Wasted time and budget.

Should've just started on Maximize Conversions right away. google already knows who converts. Someone on here pointed out that Max Clicks trains the algo to find clickers, not buyers-they're totally different audiences for mobile apps. Switched it now with a tCPA about 25% above my historical average. Let's see if it recovers.

But honestly, if the only change was the bidding strategy and I'm still at 1 install in two weeks, the issue's probably deeper. Keyword matching? I've been pulling search terms and a ton are junk. Landing page relevance? Maybe my store listing or creative set is off. and yeah, my old UAC campaigns are still running-could be cannibalising the new one. Paused them temporarily to isolate.

Also been using a tool to flag when UAC starts eating into search campaign budget. caught a bunch of crap search terms I'd missed manually. Worth it if you're juggling multiple accounts and can't babysit the search terms report every morning.

Lesson learned: trust your conversion history, don't default to clicks unless you're targeting a brand new audience or conversion event. And check your cannibalisation first.

wordweaver

Oh honey, you've been doing the clicks-first dance for two weeks? That's two weeks of training Google to find the "ooh shiny" crowd instead of people who actually open their wallets. Painful.

With 300+ conversions in the account, you should've started on Maximise Conversions from day one. Google already knows your buyer profile. What counts as a conversion here?

That "clicks first" rule? It's for fresh accounts with zero history. You're way past that. Now you've got to switch immediately. Set a target CPA about 20-30% above your historical average. Gives the algorithm some breathing room while it unlearns its click-chasing habit.

If the only change you made was the bid strategy and you've seen just one conversion in two weeks? That's not a bidding problem. Check your search terms for junk, make sure your ads actually match the landing page, or face the fact that the market or offer shifted while you were busy optimising for window shoppers.

Also, is PMax still running? If yes, pause it. It's probably eating your new search campaign for breakfast.

24fire

Conversions are real people with intent, clicks are just noise from the machine layer. In a world where attention is manufactured at scale, the only metric that still maps to human behaviour is the one that requires a decision, not a reflex.

davidsearch

1 in 4 from max clicks? That's a bloody solid CR for pest control. Most people in that vertical are seeing 10-15% at best. You've clearly done the groundwork on site UX and copy.

My concern with staying on max clicks long term is the diminishing returns. As you scale, the algo starts serving the cheapest, lowest-intent queries - those price shoppers you mentioned will balloon. What's your current daily budget? Once you push past a certain threshold, even with perfect landing pages, you start eating impression share on bottom-funnel terms.

Have you tested a switch to max conversions (even with a low target CPA) to see if you can hold that 25% CR while increasing volume? I've run similar local service accounts where max clicks got us the cheap leads, but max conversions gave us more total conversions at the same spend - because the algo went after the searchers who actually booked, not just clicked.

And honestly, if your ad copy is tight and your reviews are that good, you might be leaving money on the table not using broad match with RSAs and a solid negatives list. The "spam click" argument is often just lazy account hygiene.

emmareach

Maximize Clicks is decent for early-stage traffic and data gathering, but once you've got enough conversion history, switching to Maximize Conversions is the smarter play. I've seen too many people burn budget on clicks that never convert. A colleague once summed it up perfectly: traffic without intent is just noise. Choose based on where you are in the funnel, not on what looks good in the dashboard.

bigcoffee

Another vote for maximise conversions, but with the usual caveat: you need a decent volume of conversion events for the algorithm to do its thing.

If the data pool is thin, try running a split test with a "conversions" objective set higher up the funnel - page view, add-to-cart, etc. That often gives the algo enough signal to start optimising properly.
Budget permitting, keep the test running long enough to hit statistical significance. Otherwise you'll just be chasing noise.

I've seen this approach cut CPA by about 20% in a low-volume account where straight purchase conversions were stalling. Worth a shot if you're stuck.

miaaffiliate

Account-level history won't fully carry over to a fresh campaign. Google needs its own data to optimise properly. Stick with Maximize Clicks until you're pulling consistent traffic, then flip to Maximize Conversions. That's the play for protecting your margins while you build signals.

buggedout

Start with Maximise Conversions from day one. I've seen people set it on accounts with zero conversion history and still get results - the algorithm adapts faster than you'd expect.

Running Google Ads for service businesses in highly competitive verticals, I've found anything other than Max Conversions leaves money on the table. Here's why:

Manual CPC or Max Clicks almost always means underbidding against competitors using automated strategies
Smart Bidding learns from user behaviour even without conversion data - you'll get early signals
If the market's already aggressive, you can't afford to waste budget on irrelevant clicks

The only caveat: give it at least a few days of solid impression volume before judging performance.

dailygrind

Mate, I've run this playbook across a dozen ad accounts and the 300-conversion threshold is non-negotiable if you want the algo to actually learn. here's why:

Signal density matters - Below 300 account-level conversions in a 30-day window, the machine's just guessing. you'll see wild CPA swings. once you cross that line, it starts optimising reliably.
Switch at the right time - Don't leave conversion bidding on from day one unless you're flushing budget. Start with clicks or link clicks to build volume, then flick over once you've banked ₃₀₀ conversions. i've seen accounts drop CPA by 40% doing this.
Watch attribution windows - If your product has a longer sales cycle (like B2B SaaS or finance), those 300 need to be attributed within the same window the model uses. otherwise you're feeding it stale data.

Bottom line: hit that 300 number, then flip the switch. Anything less and you're paying for the algorithm's training time.

marcuthread

If you already have strong conversion history, I’d probably lean toward maximize conversions.

chloeest

I'd always run a split test: Max Conversions vs Clicks, pure and simple. a colleague once saw conversion rates jump nearly 40% after switching to Max Conversions, but only after letting it bake for at least three weeks. the key? Tighten audience targeting before the switch - otherwise the algorithm just spends on cheap clicks from lookalikes that don't convert. I've got a heatmap from one test that shows the difference in user behaviour crystal clear. Data before dogma

ranktracker

Only start with clicks if you're testing a fresh conversion action or going after a completely different audience. i've seen too many people optimise for clicks by default when their existing funnel already has conversion data - you end up wasting budget on irrelevant traffic. Run a few scripts on your log files first to confirm whether that new action actually mirrors the behaviour of your established cohorts. If it doesn't, clicks can be a useful starting point to generate enough volume for a proper conversion model to kick in later.

ryanseo

Switching to Maximise Conversions makes sense given that history. The tCPA buffer is smart - gives the algorithm room to breathe without strangling delivery.

Had similar cannibalisation problems a while back. Started using ChadAd.s to flag when PMax starts eating into search campaign budget. Also caught a load of junk search terms that manual reviews missed. Worth setting up if you're running multiple accounts and don't want to babysit the search terms report every morning.