This free, no obligation performance analysis will uncover:
    • Opportunities to earn more from your catalog
    • Ways to optimize performance across search, shopping, social, and discovery channels
    • Wasted ad spend and how to reduce it

3 Signs Your Approach to Automated Bidding Isn’t Working

Mike Perekupka headshotAutomated bidding comes in several forms for retail marketers. You might apply Google’s tools, work with a technology provider, or use an in-house solution. How do you know if your approach to automated bidding is or isn’t supporting your performance goals?

We’ll answer that question right here. Below is the transcript from our webinar, 3 Signs That a Marketer’s Automated Bid Approach Isn’t Working. Led by Sidecar’s Mike Perekupka, this webinar examines overlooked KPIs that reveal if your bid approach fits your goals.

Learn if your automated bidding approach is the reason for any performance loss, and whether a performance dip necessarily calls for rewriting your bid approach. Mike presented this webinar as part of the TechTalks series for RetailX 2020.

This transcript has been edited lightly for brevity and clarity.

Why Bidding Is Complex

I want to start by talking about bidding. Bidding is challenging. And I want to just level set again with why. I’m talking specifically about Google Shopping with this particular view on the screen, but it applies outwardly to other channels.

Bidding is a silent auction, and that in and of itself is tough. In Google Shopping, you don’t get things like your position. There are hacky ways to get through it. There are some metrics that you can glean your position from. You have to identify: Do my customers go to the shopping tab? Do they stay on the search engine result page?

Google introduced this concept of free shopping ads. How important are they to the business? Then there’s just the concept that it’s always changing. There’s this scale. There are people searching at different times of day, on different devices, in different locations. How do I get these strategies to work together?

This is a challenge. Most people have said, “A human shouldn’t really be doing this. We need to adopt some technology.” As you think of the space of technology solutions that are out there, there are three large buckets.

Automated Bidding Options

If we’re talking specifically about Google, Google itself has put out many automated bidding solutions. Most recently with Smart Shopping campaigns, they have a suite of smart bidding technology that includes target ROAS, maximize clicks, maximize conversions, and in the past, enhanced CPC.

These are solutions that Google has offered and encourages not even just retailers but everyone to use, because they believe they do it best. There are some perhaps scary implications of letting Google manage your bidding on a channel where you’re also advertising on, but that’s a different story.

The second bucket would be technology platforms. These are people that have built solutions that manage on top of Google to do your bidding. Lots of solutions out there. Some of them have a number of customizations, some of them don’t.

And then the last bucket would be do it yourself. Either create a system of scripts or technology yourself, or create a team that has a well-oiled-machine approach of making sure that you’re looking at the right data and making the right changes.

Each of them has positives and negatives. But the key question to set everything up for the rest of this presentation is: How do you know when it’s time to change? I want to give three signs that it actually may be your bidding solution that’s holding you back. It’s time to put it under the microscope and shake it up perhaps.

Conduct a Root Cause Analysis of Your Automated Bidding Approach

I want to walk through what’s called a root cause analysis of your automated bidding approach. I want to point your attention to three particular places that you might not otherwise look, and some trends or some metrics or some concepts underneath each one that can help you identify, “Okay, this problem actually is related to my bidding approach.”

I did my best to make this not the obvious ones, not the places that everyone goes to, because like I said, I think most people join a webinar to actually learn something that they didn’t know before.

Root Cause Analysis of Automated Bidding Approach

Metric #1: Revenue Per Click

At Sidecar, we’ve spent countless hours learning about the problems and understanding every in and out of bidding. I want to share that knowledge with you. I want to start with the concept on the left. It’s a metric that some people talk about, but it’s not on the tip of the tongue for many marketers out there. This is not a specific report that you can download, but this is a metric that I think people should be more interested in than they currently are. And it’s called revenue per click, shorthanded to RPC.

Revenue Per Click

Revenue per click is the amount of revenue that you should expect or associate with a single click in your PPC campaign. I think a lot of people very much understand conversion rates and average order values, but RPC is one that goes under the radar.

I want to walk you through why this can be valuable, and then wrap it up with a scenario that I think it’s really tied into. That’s how each of these sections will work. I’ll start with a view of data, talk about the sign that I would encourage you to take a look at, and then tie it in with a real-life scenario where this can be valuable to your business.

So let’s start with this concept of revenue per click. It is a multiplication of two very understood concepts, your average order value and your conversion rate. So getting a little bit in the weeds here, orders would cancel out, so you’re left with the revenue per click.

Revenue Per Click Calculation

Take any item or any brand or any product type that you sell in your catalog, and multiply it by an efficiency goal. I’m using an efficiency goal to be cost over revenue. If you multiply these two concepts together, mathematically it does something very unique, very informative for you. Because when you do that, the revenues do end up canceling out, again, mathematically, and you’re left with cost over click.

Why is that important? Because CPC is the name of the game in these channels that we’re talking about. This will give you the ideal CPC that you should be paying for any product, for any brand, for any product type.

Sign #1: Your Ideal CPCs Do Not Align With Your Actual CPCs

And this is what I’m going to call is my first sign. So it’s an easy calculation to grab. But if your ideal CPCs are not aligning with your actual CPCs, that is your first sign that your bidding approach may not be working.

Here’s an easy example. You can see in rows one and two, the ideal CPCs are actually greater than what I’m currently paying. So if you have room to grow your account, you should be looking for examples like this, where you’re only paying 65 cents, but your ideal CPC is greater.

Your ideal CPCs do not align with your actual CPCs.

You can go capture more impression share, go get more clicks, go bid up, if you will. You know, win those types of options. If you’re on the other side of that, your ideal CPC might be 45 cents. You’re paying too much. That’s a great indicator of it’s time to pull that back. So that is my first sign, if these two metrics are way off. And your definition of way off can change, but use the eye test. If they’re not lining up at all, I’m suggesting it as sign one.

A real-life example where I think this can be put into play is with budgets. If you are in the scenario where you’ve received incremental budget, your performance is doing well, and they gave you more money, but you go to your account and you can’t figure out where to spend it, I would encourage you to look at your RPCs and see if they line up.


A problem that I would call out is if your automated bidding approach is target ROAS. Those levers are taken away from you. Your bid column is essentially taken away from you. Google handles it. And even other things like audiences and time of day, they’re not used. Go calculate your RPCs, calculate your ideal CPC, and see how they look. So that’s sign one.

Metric #2: Keyword/Placement Report

The next metric takes a little bit of clicking around your account, but has a ton of benefit. It is your keyword/placement report. If you manage text ads, I think you’re already familiar with this report. But if you’re doing Shopping, I think there’s this misconception that you can’t actually view this report in Google Shopping, but you can.

Go to your Google Ads account, and you click through to Reports, under Basic Reports, and then Search Terms. That’s how you start. The key step here is, in the row column, add your keyword/placement as a metric on the right. And any column you’re interested in, this is going to help you evaluate.

How to Download Keyword/Placement Report

The reason I think this has gone under the radar is this used to populate as a dash-dash in the interface. Most marketers just dismissed it. But when you download it, it does populate. And now even in the UI, it does populate. It’s actually a mess. I think that’s another reason people have started to ignore this.

But it’s very valuable. It gives you the link between the search terms that your products are showing up for. It’ll say there is a search, and this is the item or the items that you’re bidding on that are showing up for that search. You’ll see a lot of equal signs. But if you understand it, you can mine through this.

Keyword/Placement Report

That first example is, you’re bidding on a product type called shoes. It matched to a search term of slippers. And it’s going to be overwhelming to look at. So what I want to do is give you a sign, give you something to potentially narrow your focus on how I think this can be useful, and then tie it back again to a real-life scenario.

Sign #2: One Bid Is Not Enough to Account for How Your Shoppers Search

If you dig through this and determine that, “Wow, my shoppers are searching very differently, and just having this one bid is not enough. I can’t account for all of these ways people are searching.” That’s sign #2 that it might be time to shake things up.

Here is example data. You can see these are all the search terms that match to a specific item. When the search term is a certain part number or model number or specific size, and you scroll over to the ROAS column, they’re the ones that tend to perform very well. That’s a suggestion that you actually might be willing to pay more than $2.10 for them. But I wish I could pull back on others.

One bid is not enough to account for how your shoppers search.

We partnered with RetailX to do a webinar about how to mine this data. I would encourage people to go check that out if this is of interest to you.

If you are seeing that this is a problem, and you’ve identified there’s a trend, I think it would be layering another strategy to your bidding approach. Put some keyword filtering in, put some of what we call search query management into it, and give every product multiple bids to account for the massive differences in how people search.

A real-life scenario where I think this comes into play is the plateau problem. Everyone experiences it. You make a change, you grow, and then you plateau. Everyone says, “What’s next? How are you going to get to the next level?” It’s very tough to answer that question.

Your keyword/placement report is where I would go first to answer “What’s next?” If you have adopted Smart Shopping as your automated bidding approach, they don’t give you that report. You cannot see these terms. That’s a red flag. They don’t give that information to you.

Performance Plateau

Metric #3: Max CPC vs. Average CPC

The third metric is the concept of comparing your max CPC to your average CPC. That means your bid versus how much you’re actually paying for each click. As a reminder, in Google Shopping, when you place a bid, you only pay one penny more than the next highest bid.

Max CPC vs. average CPC is what I would call the only mathematical view into the competition. Bidding is a silent auction, but you can, by looking at these metrics and then tying them up with some others, get a view of what the competition’s doing.

Max CPC vs. Average CPC

You can see if you’re way outbidding the competition. By raising a bid, it doesn’t stand to do much, especially when impression share starts to go above 90%. You can raise that bid all you want and it really shouldn’t have that big of an impact. When some of these numbers get very close — like 67 cents compared to 64 cents — that’s a competitive space.

Sign #3: Your Automated Bidding Strategy Can’t Keep up With the Competition

Why is this valuable? That’s the third sign. If your automated bidding strategy can’t keep up with the competition and it’s not using the right metrics, that’s my third sign that it might be time to switch up your automated bidding approach.

Here’s an example that I would give you to do, a small task. It just takes two weeks. You don’t even have to do too much. You just have to be mindful to go and check it. Don’t pick your entire catalog to do this on. Let’s focus on a particular brand or a particular product, like I’ve called out here.

Your bidding strategy can’t keep up with the competition.

For one week, monitor these metrics. Look at your bid. Look at what you’re paying for that bid, the average CPC. Understand that difference, and then compare it with the metrics I’m calling out here — impression share, the percentage of time you’re showing up. Also take a look at the benchmark CPC. This is the average bid that Google gives you that the competition is bidding on similar products. And then obviously your performance, orders.

Monitor again for a second week. In this case, the data suggests the bidding didn’t change. It’s still at $1.50. But take a look at how all the other metrics may have adjusted. The average CPC is now $1.45. So it’s now getting more competitive.

Your impression share has dropped, which means someone is doing something in the market and capturing that space from you. The benchmark bid, which goes hand-in-hand with it, has now raised. And that has caused your orders and your performance to drop. If your bidding strategy is not keeping up with that, that’s suggesting your bid should actually be higher to now compete in what is a more competitive space.

Another quick example — SKU123. In week one, max CPC is 80 cents and average CPC is 78 cents. The difference is only two cents. Again, an immediate sign of a highly competitive auction. 

The performance is very strong. But if your bidding solution is only looking at ROAS, and if say after week two it did bid it up by 10 cents, I’m still suggesting this is a sign it’s not keeping up. Take a look at what happened. It’s still competitive. Your impression share may have grown a little bit, but the benchmark CPC hasn’t changed, which means competitors aren’t changing their strategies on this particular product.

You could have taken a whole week and your performance could have gone stronger. There was a lot of potential for this product, and by not having a bidding solution that’s monitoring all this and keeping up with that, I think it might be time to say, “Bidding is the reason that my performance is not the best it can be.”

Performance Dip

Performance is going dip at certain points, and you’re going to have to try to pinpoint the problem. Depending on your expertise with this channel, you may have tons of things to check. Maybe inventory, maybe website, mobile site. I would suggest take a look at what the competition did.

In Summary: Root Cause Analysis of Your Automated Bidding Approach

If you’re asking yourself whether it’s time to make a change with your bidding strategy, there are three signs that I would call out. If your ideal CPCs are not lining up with what you’re actually paying, I would say that’s a huge sign. Your bidding technology is not working towards your goals.

If you look at your keyword/placement report and realize one bid is just not enough, I think that’s a good second sign. Last but not least is, it’s a highly competitive space, and if your bidding strategy isn’t keeping up, it’s time to make a decision.

3 Metrics That Show Your Automated Bidding Approach Is Wrong

I’m curious if there are any questions I can answer.

Question: I’m using target ROAS. Can I still view the keyword/placement report and act on it?

We get that question a lot. Yes, that report is still available. Within Google you can still download it. But what you can’t do so much is act upon it, because you don’t have that last lever of bidding. But I would still encourage you, and if you use target ROAS, to download that report and see what you can learn from it.

Question: Is there a certain time period you recommend analyzing when you are evaluating these metrics?

If you’re looking at that keyword/placement report, don’t go below 90 days, unless you have just a ton of traffic. You want to get some significant data to identify those trends. But the other metrics, I think it’s fine to look weekly or at least biweekly. Just because if you’re not, I think you’re also not keeping up with the competition.

Question: It seemed that a lot of the signs today were oriented around Google Ads. Do these or similar signs apply to other channels, like Facebook and Amazon?

Yes these concepts apply. I would say they apply directly to Amazon because Amazon is so similar in many ways. Facebook is a little different because you bid on audiences, and actually Facebook discourages you from making too many changes. They throw your account into this idea of learning mode. But it’s applicable there in different ways.

Question: Would you trust target ROAS? We attempted to test it, and bids fluctuated dramatically, causing CPC to skyrocket and ROAS actually dipped.

Yes and no. I trust it to do certain things. But I would say if you have enough data and you have the right technology, I don’t trust it to do most things. It’s not very good at hitting a certain budget. It is very good at spending it. You say it caused CPC to skyrocket. That I’m not surprised by.

There’s a time and a place for tROAS. If you’re time strapped and that’s the only thing you can do, I think it does a decent job. When you start getting a good amount of data or if you have an advanced strategy or nuanced parts of your business, there are better options to be honest.

Question: Do all these metrics have to be going awry to justify changing your automated bidding approach, or does just one metric have to be going awry?

I don’t think all of them have to be going awry. I think if they are all going awry, I’d say that’s a huge sign. I think if any particular one of them is going awry, I would at least take the time to test. If you can create a test where you can learn and measure, even just one or two of these signs might be enough to see if it’s worth changing your bidding solution.

Metricool tag