Google's Smart Bidding Change: Why Your ROAS Is About to Drop

Episode Transcript

00:00:00 - 00:00:22
Mike: Welcome to another episode of Growing Ecommerce. I’m one of your hosts, Mike Ryan. And with me, as always, Chris.

Chris: Good to be back.

Mike: I should say, almost as always; you weren’t with us for the last one, but you did a fantastic job even without me somehow. Wow, what a surprise. So today we have a very exciting episode for you.

00:00:22 - 00:00:48
Mike: First off, at the time of recording on July 3rd, we’re in the middle of the World Cup as well. Chris is combining these and wearing his US jersey. Right? Yeah. We’ll make sure to cut to you on the camera there. And what could possibly be more American than the raw capitalistic energy of Google bidding? Nothing.

00:00:48 - 00:01:22
Chris: Yes, indeed. I think let’s start.

Mike: Yes. So a couple of weeks back, Google wrote a blog post talking about some changes coming to smart bidding. They talked about a new feature called Smart Bidding Exploration, which we’ll probably touch on this episode as well. But what sounded like a very dry administrative topic was “changes to smart bidding beginning August 17th.”

00:01:22 - 00:01:48
Chris: I’m sure it was the least-read part of that article, and they were great at hiding it.

Mike: Yeah, just be very defensive about it. Well, they do have a tool coming out, so they will be a bit more proactive here. But that tool, I think, is going to have a very specific purpose. I haven’t seen a lot of deep discussion yet about these changes or what they mean.

00:01:48 - 00:02:12
Chris: No, nothing on the client side either. But I think, without being too bullish or bearish, it’s a question of perspective. There will be discussions going on very, very soon. But before we jump into what it actually means and the positives and negatives, can you explain shortly what it is?

00:02:12 - 00:02:38
Mike: Sure, let’s break it down quick. Google’s framing this very much around campaigns that are limited by budget. As we go on, I think it’s much broader than that. My understanding is that this is a systemic change that’s going to affect a lot more campaigns, but their initial framing is about “limited by budget” campaigns. Chris, does Google like it when a campaign is limited by budget, in your experience?

00:02:38 - 00:02:58
Chris: I have to think about that… No, of course not. We’re talking about a company with a rather capitalistic approach to business. So, “limited by budget” is, by design, not good.

00:02:58 - 00:03:25
Mike: Yeah. And what’s going to happen here is they’re saying that it is often the case that campaigns limited by budget might have a target ROAS of five, but the campaign is actually delivering a ROAS of ten. They have the feeling—and I think there’s some truth to this—that some people are just comfortable in that situation. They’re using cost as a control and are happy with that buffer. If they tried to scale that campaign, they know their ROAS is going to go down, there’s going to be volatility, and there’s an uncertain upside potential. They’re leaving revenue on the table, but they’re very comfortable in this position.

00:03:25 - 00:04:18
Chris: Also, if we take Google’s perspective—because we are in our bubble with our clients and how they tend to spend—I’m willing to bet when Google looks at the entire landscape, there are probably so many campaigns, particularly in the SMB segment, that are in this exact situation. I think you did an analysis over our MCC where you analyzed how many campaigns are over- or under-delivering in terms of ROAS.

00:04:18 - 00:04:43
Mike: Yes, and there is quite a substantial part of the existing campaigns we can look at—we’re talking about thousands of campaigns—which are over-delivering, basically delivering a ROAS above the actual ROAS target.

00:04:43 - 00:05:17
Mike: Definitely. I just want to table that for right now; we’ll come back to it. Because what Google says is going to happen starting August 17th is that those campaigns limited by budget and over-delivering efficiency—that behavior is going to change. They’ll start to trend toward the target ROAS. So if you have a target ROAS of five and you’re actually getting ten, beginning August 17th, that campaign is going to look more like a ROAS of five. On paper, I would say Google is just doing their job; ROAS adherence could even be a positive thing.

00:05:17 - 00:05:44
Chris: Yeah, I mean, Google has no obligation to give you a ROAS of ten if you’re asking for five.

Mike: If I want my ROAS at five, which is the target I put on my campaign, and I get a ten, they are over-delivering. But there’s a huge “but.”

00:05:44 - 00:06:03
Chris: We are talking about campaigns which are limited by budget. Let’s say I’m a retailer now; I have one of my most successful campaigns with a ROAS target of five. It’s delivering a constant eight to ten. I got accustomed to it; I love it. But the campaign is limited by budget. What you’re telling me and the audience now is that Google will revert back to what I actually set with regards to my ROAS target.

00:06:03 - 00:06:26
Chris: So from ten, it will go down to five, and the budget still remains limited. That’s a fix, right? I want my limited budget. This means, for example, maybe it’s at 90% of the available budget, so this becomes a fixed constant.

00:06:26 - 00:06:49
Chris: Hold on now, because this is where we’re jumping right in. What does that mean? It means: how can Google force my ROAS down to five from ten without increasing the volume? There are just two ways to achieve that. Focus on clicks with less expected value or a lower conversion rate.

00:06:49 - 00:07:15
Mike: Holy shit. So they are deliberately looking for a less efficient conversion?

Chris: Ultimately. The other side of the equation is they can go into more competitive auctions with higher CPCs.

00:07:15 - 00:07:43
Mike: Higher CPCs? Yes.

Chris: Which I think is the preferred one for Google. You can’t even say “because the budget was…” what they actually want is to motivate people to increase or open the budget limits.

00:07:43 - 00:08:14
Mike: Yeah, but that’s fine for me. Going back to these two ways they can achieve this forcing function to lower the ROAS—honestly, this is crazy, mate. I know they don’t owe me if I put a ROAS target of five, but still, it’s crazy. They are deliberately looking for less efficient conversions.

00:08:14 - 00:08:43
Mike: It does come down to this idea of they don’t owe you. To break down what they’re offering here: first off, I think this is a “carrot and stick” approach. If you want a horse to do something, you can either offer a nice crunchy carrot or you can hit it with a stick. And by the way, Chris is a huge animal rights fan, so it should be the carrot.

00:08:43 - 00:09:03
Chris: Is this a carrot here?

Mike: Well, what Google says will happen—because they know people are in this comfortable situation of a limited budget and over-delivering ROAS—is their offer of “smoother scaling” if you do open your budget. That way, you can scale more easily. I’ve seen slides where they show someone 5x-ing from 200 to 1,000, and the ROAS remains perfectly constant at ten. I think these kinds of explanations are unhelpful because that’s not credible.

00:09:03 - 00:09:29
Mike: There’s this little thing called the law of diminishing returns. Notwithstanding that, they’re offering you smoother scaling—that’s the carrot. The stick is that your performance will not be as good if you stay capped. If you’re asking for five, they don’t owe you ten. They’re saying if you want that 10, then start asking for 10.

00:09:29 - 00:09:53
Chris: Fair enough, because otherwise there’s someone out there who’s trying to scale who is, to be honest, a more attractive customer for Google. They need what you have and what you maybe didn’t even want in the first place.

00:09:53 - 00:10:25
Chris: Potentially. From the perspective of Google, it’s a bold, smart move. It will probably motivate people to really open the budget. If they push my ROAS down from 10 to 5 and I got accustomed to eight, I probably need that eight for my overall ROAS on the account level. I’ll probably do something about it. I’ll either increase the ROAS target or accept the lower ROAS but then try the scaling thing, which is the big promise behind this feature. It’s very shrewd. Advertisers might not like it, but from a business standpoint, it makes a lot of sense.

00:10:25 - 00:10:50
Mike: We can look now at the bigger picture with other campaigns. But what did you want to say?

00:10:50 - 00:11:32
Chris: I wanted to refer to how much slack you have here. I think they have a lot of bending room with this diminishing returns curve.

Mike: I agree. Let’s get into this topic of smoother scaling. How should this possibly work? Again, I think some of their marketing materials are not realistic or helpful, but I do believe it’s possible for them to offer smoother scaling.

00:11:32 - 00:11:57
Mike: The question is: how? There’s the law of diminishing returns; it’s a universal law. In marketing, we talk about the saturation point. But that curve is uncertain and flexible; you can bend and shape it. There are things you as an advertiser can do, like improving your on-site conversion rate, but there are also things Google can do at a system level.

00:11:57 - 00:12:43
Mike: As you mentioned earlier, Chris, what we’ve observed time and again is that smart bidding is fundamentally conservative. I’ll put a slide on screen for people watching. We find that in campaigns that don’t have a lot of monthly conversions, there’s a huge amount of volatility between whether you’re going to be on target, above, or below.

00:12:43 - 00:13:09
Mike: As you start to pack on monthly conversions and the algorithm has more data, the percentage of campaigns underperforming the ROAS target diminishes and eventually goes close to zero. The number of campaigns on target increases. But what’s truly fascinating is that what never changes is the percentage of campaigns that over-perform the ROAS target. It’s about one in four—roughly 20% to 25%.

00:13:09 - 00:13:38
Chris: This is significant. This is the slack Google is looking at and saying, “Hold on, this is money left on the table.” That’s why the wiggle room they have in this diminishing returns curve is serious. This feature might work.

00:13:38 - 00:14:07
Mike: Google says it will. And by the way, they already rolled it out for display campaigns. I think they used that as a guinea pig because they’re killing display campaigns anyway. Although they’re focused on “limited by budget,” I think that’s the largest population of campaigns affected. From a sales perspective, they want to uncap those budgets.

00:14:07 - 00:14:28
Mike: Logically, the technology they’re describing doesn’t stop working as soon as your campaign’s not limited by budget. They’re saying you can keep spending more and manage to the curve of diminishing returns; you’re going to get this smoother scaling. That means better overall ROAS or target CPA fidelity.

00:14:28 - 00:14:52
Mike: We contend that the percentage of campaigns that over-deliver will probably shrink, and the percentage on target will grow. From a bidding standpoint, these limited budget campaigns will start facing higher or lower order values, and it ultimately motivates you to spend as much as possible.

00:14:52 - 00:15:28
Chris: Look, I don’t want to say that we were right again, but this is something we’ve been talking with our clients for years—that Google is just aiming at this “average ROAS” idea. You have over-deliveries and under-deliveries. Even if you look at one campaign, products are over- and under-shooting all the time. Our claim was always that if you have a campaign over-delivering in terms of ROAS, that’s not necessarily a good thing because you leave market on the table.

00:15:28 - 00:15:56
Mike: Exactly. From that perspective, Google is doing the right thing. They motivate you to think about whether you want to capitalize on this market. If you want to, you have to put money on the table. That’s the game. Honestly, Mike, I think it’s shrewd and bold. It’s bitter medicine, but it might help.

00:15:56 - 00:16:23
Chris: The one thing I’m not certain about—and I know our beloved clients—is that sometimes there is just no budget available to scale, even if it’s “smooth scaling.” If I need more budget and I don’t have it, it might hurt players who are truly limited by budget but got accustomed to this over-delivering ROAS.

00:16:23 - 00:16:47
Mike: This might be a real problem. I’ve already been in my first client calls about this because August 17th is not far. Again, we always talk about how Google doesn’t know what they don’t know. Plenty of marketers don’t really care that much about the platform data; they’re looking at their own back end, attribution model, or marketing mix modeling.

00:16:47 - 00:17:08
Mike: People are moving toward more advanced measurement. For one client using Adobe, that’s their real source of truth. In Google, the over-delivering ROAS they’re currently achieving actually corresponds to break-even in their back end. What’s going to happen when Google starts trending that toward the target?

00:17:08 - 00:17:42
Chris: They’ve learned they can ask for this ROAS target at this budget and get this actual ROAS, which corresponds to their back-end ROAS. It’s tangled and complicated, but it’s working for them and it’s about to break. The understanding of my ROAS targets—in terms of what I have to set to achieve a ROAS that meets profitability goals in the back end—might be broken.

00:17:43 - 00:18:03
Mike: What is the workaround? You have to be way smarter with your product segmentations and campaign structures, because the campaigns will probably hit the ROAS target you’re setting there.

Chris: I think so. But these are two massive issues for the online retailer.

00:18:03 - 00:18:27
Mike: I don’t want to scare people with this overly. Google is not presenting this as a cliff. There’s a date where the system will be activated, and then they say campaigns will trend toward this result over time. You cannot guarantee or depend on the results you’ve been getting.

00:18:27 - 00:19:00
Mike: Also, by the time this episode airs—because on July 6th, they’ll be offering a tool called the Bid Target Adjustment Tool.

Chris: Be more creative, Google. There’s no “Max” and no “AI.” Can we please call it the AI Bid Target Max Adjustment Max Tool?

00:19:01 - 00:19:25
Mike: Sorry, Google. I guess that tool is going to paint a certain picture that you should probably increase your budget.

Chris: Sure thing. From a microeconomics perspective, this makes sense—the curve of diminishing returns. You have market; spend it. They offer a tool. But we both know that’s not always the reality. A lot of clients don’t have the budget, or they need this precise ROAS to meet their goals.

00:19:25 - 00:19:46
Mike: It’s very broadly aligned with other activities Google is taking. You can read the theme here: “demand-led growth.” They announced at GML that there will be a new feature called demand-led budget pacing. They have this whole smart bidding exploration arm coming to more campaign types. They’re building features like “promo mode” coming this holiday season.

00:19:46 - 00:20:06
Mike: They also announced flight-based budgeting. Everything is pointed in one direction: automation. In the past, we talked about automated bids, targeting, placements, and creatives. The things left were your goal and your budget. That pillar is increasingly coming into the platform.

00:20:06 - 00:20:29
Chris: I don’t know if it’s a coincidence, but since they were behind for basically the first time in history with regards to ads budget, Google has to do something about it. In my world, I’m an opportunity seeker. If you have campaigns that over-deliver, you have market there; think about how to activate it. That has always been my claim. The reality is it won’t be a fit for all retailers or DTC brands, and this might cause friction.

00:20:29 - 00:21:03
Mike: It’s easy to say, “spend more,” but it will be more important to think about your product segmentation strategy and campaign landscape. The ROAS you set will matter even more in the future because if scaling is getting easier in Google’s platform, how much can your business actually scale profitably? Where is the point on your break-even? What is the global optimum across all these campaigns you want to achieve on an account level?

00:21:03 - 00:21:26
Chris: It’s a hell of a thing. Is it a feature?

Mike: Well, no, the feature is smart bidding, and this is an update to that feature.

00:21:26 - 00:21:48
Chris: Yeah, that’s a hell of a thing.

Mike: Yes. You know what else is a hell of a thing? The U.S. boys. July 4th is coming. The soccer team—or football team—is playing amazing football. I predict they will beat Belgium.

00:21:48 - 00:22:12
Chris: Well, we know how that goes by the time this is aired!

Mike: You’re going to have to school me on this because, like the true American that I am, I don’t really care about soccer. Why is it called football? You play 90% of the time, but the foot touches the ball very limited. Anyhow, the U.S. boys are a great team. Austria… we followed our butts off, but Spain was just too good.

00:22:12 - 00:22:56
Mike: I knew this moment would come. I was going to take out my watch, put 15 seconds on the clock, and say, “give me your pitch.”

Chris: No, I don’t want to be too mean. Life goes on.

00:22:56 - 00:23:21
Mike: Exactly. I think this is it, right? Thank you, Chris. Enjoy the weekend.

Chris: Likewise, I’ll try.

00:23:21 - 00:24:08
Mike: Thanks to everyone for tuning in to another episode of Growing Ecommerce, brought to you by Smarter Ecommerce. You can learn more at [link removed]. If you leave us comments on YouTube or Spotify, or give us a shout on social media, we really appreciate it. We’ll see you next time. Bye-bye.