PPCChat | What Are We Actually Optimizing For Now in PPC? | 21st July 2026

This week’s PPCChat discussion focused on what experts are optimising in PPC. Does optimization vary by platforms or clients? Have their targeting strategies changed in the last year? Is AI playing a role in your optimizing or targeting strategy? Host Julie F Bacchini sought answers to these questions and more.

Q1: What are you primarily optimizing for in your campaigns these days? Does it vary at all by platform or client/brand?

So obviously, I am still optimizing for conversions (in my case leads). @NeptuneMoon

Has always varied by the client’s goals. @revaminkoff

I do find the AI of it all is impacting a lot though @NeptuneMoon

It does vary, but ultimately it’s conversions or engagement. @JuliaVyse

Subsequent questions will dive into how we are thinking about optimizations etc. @NeptuneMoon

I only run Google Ads, and only for lead generation. For most of my clients I’m optimising for a qualified lead conversion. Either hand-marked, or using a proxy like booking an appointment or scheduling a visit by an estimator @Pete_Bowen

conversions, generally sales and form leads.  I occasionally do other types for demand gen-like campaigns, but when it comes down to it- I’m trying to optimize for cost per conversion, value/conv, and ROAS @ynotweb

And to clarify, it is usually a cost per lead/acquisition that is the primary metric the client is interested in @NeptuneMoon

We do a lot of ecom, purchase is what we optimize towards. We also look at conversion rate on site when doing UX/UI/Mobile audits to make the site better before spending more on ads. @duanebrown

My engagement clients are interested in deep interest, like video plays and view through rates @JuliaVyse

Q2: Has your thought process on what you are trying to optimize for in your PPC strategy and management changed in the last year? If so, how and if not, why not?

The closer to the actual goal/revenue we can get for the client, the better. @revaminkoff

This question is really interesting to me, because while I feel that what we are truly optimizing for (conversions at a particular cost) hasn’t changed, the whole ecosystem around how things works on ad platforms has. @NeptuneMoon

no, though I anticipate it will in the NEXT year. @ynotweb

@revaminkoff you’ll have to wait till August 17th @Priteshpatel

We’ve been dealing with a lot of concern over zero click, so a lot more on ad copy and connecting youtube and search ads @JuliaVyse

One big example is how much the system chooses “winners” in campaigns now vs. how that used to work. It is so much harder to test ideas or approaches than it once was. @NeptuneMoon

I’ve been trying to optimise on search discovery recently but my Ai algo gives up too easily and quickly. @Priteshpatel

We had to launch two new creatives at the exact same time to get a read on which was better. When we ran it against the old creative, it barely served. @revaminkoff

And, the system chooses its “winner” really fast @NeptuneMoon

Algo is like a little kid. Throw a sweet and chases that sweet instead of searching for the bag of sweets. @Priteshpatel

My theory on this is that once the system finds something that works relative to the bidding parameters, it wants to ride that horse for as long is it possibly can before it explores other options or avenues @NeptuneMoon

Talk more about profit and looking at things beyond the ad account. We don’t sell items with paid ads if the client would lose money because margins are too small. @duanebrown

@duanebrown even tho loss leaders “can” fuel the “Ai algo”?  @Priteshpatel

Once you train the AI to chase easy things (loss leaders) it has a hard time shifting away from that. It ALWAYS wants the easy win/low hanging fruit first. @NeptuneMoon

On tricky B2B accounts (with $200 clicks), I find myself turning more and more to my broader marketing background. Instead of search, demand gen to drive newsletter signups, that sort of thinking. @InquisitiveMarketer

That is why you have to be careful with micro or stepped conversion actions. You must include values that have a strong relative meaning to the value of the actions! @NeptuneMoon

Not sure if the change has been in the last year, but I’ve definitely started thinking a lot more in terms of optimsing for qualified leads rather than just raw leads. Put a lot of work into getting my clients into a position where we can do that @Pete_Bowen

I also think that we have to think about optimizing for the way the platforms work now too, way more than we used to. @NeptuneMoon

@Priteshpatel Our goal is to make money and doing profitable. AI algo can go to hell @duanebrown

@NeptuneMoon for sure. @duanebrown

Previously, the platform was kind of a static tool we used to put together ads, choose detailed targeting and show ads to potential customers @NeptuneMoon

Now? It is making all kinds of decisions and judgements on both targeting and the content of ads shown. Which is very, very different. You have to plan for how Google thinks in addition to what the client/brand wants and how you think you should go about presenting it and to whom. Ad schedules don’t even function like they used to with regard to budgets @NeptuneMoon

And Google is trained on general behaviours. god help the niche offers out there @JuliaVyse

@JuliaVyse B2B is getting more and more difficult because of this @Pete_Bowen

I’ll take that one step further @JuliaVyse Google is trained primarily on ecommerce. And it thinks like everything is ecommerce. Which is great if you are selling ecommerce things, but if you’re not it’s levels harder. @NeptuneMoon

@NeptuneMoon I don’t think you need a Q3 – this one is a bonanza @Priteshpatel

@JuliaVyse right. @ynotweb

Yes! Even auto is having a tricky time, which is neither niche, nor small @JuliaVyse

Can attest. Even clients that are SELLING but not selling things Google considers PRODUCTS (i.e. can’t go in Merchant Center) is at least 3 times harder @ynotweb

We have to remember too that today’s AI does not think, it pattern matches. And you have to choose those patterns to match to as a platform. Which serves some advertisers much better than others. This is true across all platforms too @NeptuneMoon

Julie, this reminds me of when credit card companies would freeze my card once a year because my usage “did not fit a pattern”, but it was literally something I did EVERY year. Just only one month out of the year. Every company, every year. They only considered 6 months at a time. I am currently seeing conversion events thrown to “unverified” if it hasn’t happened in a week or two. @ynotweb

Google’s upped the fear factor in the notifications. Even stuff that’s normal eg a day without a conversion on a small account now gets flagged as an error. This has added a bunch of unecessary work and explaning @Pete_Bowen

Q3: Is AI playing a role in your optimizing or targeting strategy or priorities? If so, how?

My feelings on AI in general are well documented, but… You can’t just ignore how AI is being used within platforms now. I shared the tweet that got me thinking about this topic up channel and it is true that we have to think about how people will potentially react/interact with our ads and landing pages but not also very much how platforms’ AI will too. And does that create any kind of disconnect or disjointedness in what we need to input to get the output we want? @NeptuneMoon

I think the fact that disabling auto generated stuff is a common practice says it all. I’ve seen all Google Ai generated headlines perform the worst against my own headlines using my brain. @Priteshpatel

The only way it plays a role is the ways that I cannot control. I choose an audience, that audience is built by their AI. @ynotweb

And we have not even gotten into the whole concept of agentic anything. How will we and platforms optimize for that if it actually catches on? Will platforms know it is an agent interacting and make different choices in what it does or will everything be muddy and people and agents end up seeing the same thing despite being very different? @NeptuneMoon

SO MANY QUESTIONS REMAIN @NeptuneMoon

I have thoughts on agentic, but will return this pm @JuliaVyse

Does automated bidding count as AI? If so then yes. Automated copy, not yet. @Pete_Bowen

Q4: Have your targeting strategies changed in the last year? What are you targeting/how are you targeting your campaigns to try to reach your best audiences? Does this vary by platform?

I feel like I have to target more broadly than I did in the past year to get a fraction of the results. @IamNextSTEPH

Well all of the platforms clearly want us to just target broadly and let them decide who should see the ads @NeptuneMoon

I have gotten way more into creating my own audiences in Meta and GA4 in the last two years. Even if I then add a layer of Lookalike, I want it based on something tangible like buying, visiting a specific page, etc. @ynotweb

Are you using exclusions differently? Platforms also really, really want your first party data too @NeptuneMoon

I am using exclusions more as well.  Remarketing feels like an entirely different game now than before. @ynotweb

Broad targeting is great when it works. Launched a campaign and immediately got $30 clicks that converted. What does the AI do? Start sending me nothing but $100 clicks that bounce after hitting the landing page. @InquisitiveMarketer

@InquisitiveMarketer this is why I have added expirations on many campaigns, regardless of early performance. @ynotweb

People I came with complaints today. What does it mean when you upload a 5000 person list, google says they’re over 80% matched, then creates an audience with like 400 people in it and keeps telling you “your audience is too small” @InquisitiveMarketer

I think getting experienced advertising brands to really understand that learning periods and the AI taking expensive swings is just part of how things work is a big piece of what I have to talk about lately too. Having conversations about the time periods we look at and sort of ignoring a super high priced click like it is the Russian figure skating judge, is par for the course now. @NeptuneMoon

@ynotweb do you mean shortening the attribution window, or disavowing the data for a certain period, or …? @InquisitiveMarketer

For Google Ads I’ve changed to mostly using broad keywords because exact match hardly seem to serve unless there is a broad match in the ad group. I’ve been reasonably happy with the results. @Pete_Bowen

@InquisitiveMarketer I mean making a duplicate and scrapping the original. This is more prevalent in Meta, but Google definitely has a timeframe as well- dependent on whether its an ecom, lead gen or other type of business that is advertising. @ynotweb

Q5: Do clients or stakeholders want you to optimize toward different things or manage optimization or targeting differently lately? If so, how?

No. I find most small business clients are at least 2-3 years behind trends in their knowledge. I’d more likely have someone come to me with something that was important in 2024 than someone who wants to make changes based on AI. @ynotweb

I think it is more generally along the lines of “why doesn’t this work anymore” or “can’t we just (something that used to work great and now does not).” @NeptuneMoon

My clients are aware that they’re not getting organic clicks anymore because of AI overviews etc., even for page 1 results, and that they’re getting corralled by the system into spending more ad money. @InquisitiveMarketer

PPCChat Participants

• Julie F Bacchini @NeptuneMoon
• Reva Minkoff @revaminkoff
• Julia Vyse @JuliaVyse
• Peter Bowen @Pete_Bowen
• Kammy Caruss @ynotweb
• Duane Brown @duanebrown
• Pritesh Patel @Priteshpatel
• Aldwin @InquisitiveMarketer
• Steph Cockerl @IamNextSTEPH

Related Links

• AI Overviews Are Costing Publishers~40% of Their Clicks — Findings From A Field Experiment
• The PPC Keyword Balancing Act: Finding the Right Keyword Count for Campaigns, Ad Groups, and Accounts
• Google’s 42 Lead Gen Launches: What Marketers Need to Know in 2026?