Episode 7 | August 1st, 2023

Google Ads Automated Bidding & Machine Learning Strategies

A2O Digital touches on the world of AI and how it relates to Google Ads and machine learning. This is a part 2 to Episode 6 Intro to Machine Learning. We dive into the most common automated bidding strategies used in the home service business. Whether you're looking to maximize more conversions and cut down costs, our team explains the pros and cons for each strategy.

Tim 00:19

Hello, my name is Tim Coleman and I am your host. I'm also the managing partner of A2O Digital a full service digital marketing agency working exclusively with home service businesses. In today's podcast, we will continue our discussion from the last episode, when we started to talk about Google's automated bidding strategies for businesses in the home service industry. If you're just tuning in, it might be a good idea to save this episode and go back to listen to part one of this. Again, our guests are Ashley Bencsko and Elainey Yerger, both members of the A2O digital team. Ashley has been working in the digital marketing space since 2008, and Elainey since 2018. Both have spent a significant portion of their time on paid search and more specifically Google ads. Ashley and Elainey, welcome. Good to have you.

Ashley 01:10

Thanks for having us back.

Tim 01:12

All right, so last time, we talked a lot about Google's automated bidding strategy, and kind of what it was and we talked about how you can leverage AI and leverage Google's machine learning with the Google ads platform, but we didn't get into the specific tactics Google allows you to kind of leverage in and how each one of those works. And I think I think there's there's kind of a bunch of these, there's like, you know, six different strategies that Google will allow you to kind of use, why don't we just start by talking about what they are. Elainey and Ashley, can you just get into a little bit about each one of these things in and as they talk about each one of these things you can imagine it almost being like a radio dot in the Google platform as you're going through this kind of deciding what strategy and machine learning do I want to employ today. So let's get started and just talk a little about the different tactics that are available to us.

Elainey 02:24

Sure. So Tim, you mentioned just now that there's six, there's definitely more outside of that. But there's around five or six that you hear most common in the home service industry. So those are really going to be the ones that we focus around. Because there are definitely ones for E commerce and other types of verticals. But for home service industry, these are the most common. So the one that we see utilized most is maximize conversions. In this strategy, we give Google a daily budget, and then they'll get as many conversions as possible for that amount. If you remember from last episode, a conversion is kind of a lead, so they're synonyms. So is getting as many leads as possible for that budget.

Tim 03:08

So if I tell you I want phone calls and people to maybe contact forms, and tell you all spend $100 a day, then Google is going to kind of go get me as many of those conversions or leads as they can for 100 bucks.

Elainey 03:24

Correct. And so you the Google will set the bid for each, each auction each search that a customer does,to get as many leads as possible given your budget. So they may bid more aggressively on a searcher that maybe leans more towards your ideal clientele. And maybe they will be less aggressively or not bid at all, on a searcher who is less likely to click and less likely to convert

Tim 03:53

Cool got it and that's called maximize conversions. What's next?

Elainey 03:56

The next step on maximize conversions is target cost per action, or TCPA. It's an extension of Max conversion. So it's the same mindset of you set a budget and they'll get you as many conversions as possible for that budget. On TCPA, you put a limit for how much you would spend per conversion or per lead. So you could say get as many conversions as possible, as long as they're not more than $100 per conversion. So I kind of think about it as like a leash. You know, Google will kind of go run as far as it wants, but only as long as you kind of set that leash

Tim 04:33

Okay, what's next that's target cost per action?

Elainey 04:37

Yup. A similar flip on this is target ROAs or target return on adspend. This strategy uses their AI bidding in conjunction with your revenue information. It has to be reported back so they kind of tie in how much you're making on the job to get you as much return on your ad spend as possible. And then Google will adjust your bids on each searcher to maximize your return on that adspend.

Tim 05:14

Okay, so here, Google is going to use revenue, we'll talk a little bit about how that magic happens. But just for now, realize Google wants to literally use the revenue you're getting on different jobs and keywords that you're bidding on to complete the circuit on that one. So that's return on adspend. We'll talk more about that in a little while. So don't worry if that's a little bit fuzzy right now, what's next?

Elainey 05:39

Maximize clicks, so similar to maximize conversions, where you set a budget, and then Google does as much as it can to maximize that budget, but instead of going after conversions, it's going after clicks. So it's trying to get you as many clicks on your ads as possible.

Tim 05:56

Okay, collect your like, I don't know what they're like, you know, okay, maximize clicks, that doesn't sound like a strategy, that's probably going to work too well, for some of our customers. Let's talk about the next one.

Elainey 06:09

Target impression share. So this one's a little nuanced, kind of probably a little buzzwords here. So your impressions, share percentage is typically the percentage of time that your ad will show in comparison to how many searchers or searches that have been done for what you're advertising on. So how often is your ad showing in row in relation to all the search that's happening? And then on top of that, it could it can go further to what, how often you're at the top of the page. So a lot of people like your target impression share, they equate it to top of page, and they want to say, well, I want to be the first ad on every single search 100% of the time. So you may see target impression share, it's really a good metric to use if you're trying to bid for your ad to be placed in a certain space on the search results page. Consistently.

Tim 07:05

Okay. So that's like a quick definition of just to get a little bit familiar with some of the strategies Google allows us to use. Let's get into each one, a little bit deeper now, let's start with Max conversions. I know that that's one that we use a lot for our home service customers. Let's get into the pros and cons of Max conversions a little bit. Ash do you want to start that conversation?

Ashley 07:40

Yeah, definitely. So the first three that Elainey talked about really are targeted towards conversions, and I didn't think of the last two is more about visibility, just like high level kind of breaking those down. But let's go one by one. So maximize conversions, we found to be a really good starting place with machine learning. So you don't have to do a whole lot of like setup or defining to get this started, you really need to decide what conversions are you tracking, get them set up properly in the Google Ads system, and then you can utilize a strategy like pretty much right away. As Elainey said, you're going to give a budget, and it's going to try and get as many conversions as it can as possible. So pros are it's easy to get turned on. Pretty straightforward, you don't have to set a lot of parameters. Cons though, are that by definition, this strategy will spend all the budget so if you are guessing at budget or way overshoot your budget, what tends to happen is you end up with a lot of inefficiencies, and your cost per Conversions can be relatively high. So let's say in theory, there's $500 worth of available traffic for your business out there on Google. But you set your budget at 1000. Google says no problem, we can spend all this right, but they start going after stuff that maybe is not quite as good of a fit for conversion. So you'll often find that you hit a cap on your conversion volume, but you don't have a cap on your cost per conversion. And that just keeps rising and rising and rising while your conversions kind of flatline. So like it's a really good starting point. And if you're good about setting limitations on your budget, it can be a really great place to actually stay for a while too, we've had a lot of accounts users for an extended period of time extremely successfully, and grow their volume of conversions and decrease their cost per conversion when managed properly. So great strategy, but you got to be really careful not to let it kind of get out of control. The next one was target cost per action, and Elainey mentioned this kind of layers on top of maximize conversions. So it's putting that leash on and telling Google I'm only willing to spend X amount on average per conversion, it means that you may get some conversions that cost a little more than that. And you may get some conversions that cost a little bit less than that. But overall average is at the number that you provide. Pros are that you're telling Google like I'm only willing to spend $100, or I'm only willing to spend $50, or I'm only willing to spend $200, and they do generally do a pretty good job at saying at least within those confines without going over, you will find sometimes they actually go under which most people don't complain about if they're getting conversions for less than they're willing to pay. So those restrictions are really helpful and a pro however, because you are limiting the machine by setting a maximum cost per action, you can lose out on conversions if you're not careful at managing the strategy. So over time, for almost every paid account we've ever worked on, right costs rise, whether it's just like the natural progression of things. It can be a variety of different factors. But those are going to go up over time, you'll also see on a regular basis that seasonality can cause your cost per action to rise or fall. So what we typically see is during very busy periods, the amount of cost to get a conversion often goes down. When it's a very slow period, we find the inverse of that right, so the cost per conversion generally goes up. If you keep your cost per conversion, or your cost per action that you've set with the strategy at what it is during a busy period, then it starts to slow down. And it costs a little more to get that conversion, but you don't adjust your strategy, what typically ends up happening is you just start losing volume and conversion. So maybe you're getting 100 conversions at $100 per conversion, now during slow period across $120. So you can only get 80 conversions at the $100 cost per conversion that you set our target CPA. So you really have to be careful to watch this. Essentially, Google treats your CPA as a bid in a different format than we're used to with manual bidding where you say I'm willing to spend this much per click, but you have to think about it as a bid, when your bid gets too low, you may start to lose out on some of your conversion.

Elainey 11:51

So you're really bidding for the conversion. So instead of bidding for the click, you're really bidding for that lead so supply and demand. Just like Ashley said, if it is more expensive to go get that you're gonna have to spend more, you have to bid more to get that.

Ashley 12:07

And I should mention that with target cost per acquisition, I think that it's a business owners desire to decide what their cost per conversion should be. So like, you know, $100 sounds great, let's do it. But if if the economy dictates that it's $150 per conversion, setting your target CPA at 100 doesn't mean that you've just lowered how much it costs to get the conversion, it just means that they'll get as many as they can for that. So that's where the maximize conversions into a target CPA strategy like using that path is helpful because maximize conversions will help you understand the market and define how much a conversion will typically cost. And then target CPA can be used to like keep your costs at a certain level or your cost per conversion at a certain level. Sometimes you can inch it down a little bit with Target CPA too. But if you just arbitrarily pick a number and hope it works out, you may not pick the right number target CPA, so you want to be careful to pick realistic numbers with CPA to and not go go for what feels the best.

Tim 13:15

The next one was return on adspend. This one seems a little bit futuristic. Because here we're going to in order to get the revenue into Google, we're gonna have to tie it to the CRM right.

Ashley 13:28

So yeah, so I think the best way to use the strategy target ROAs is to use the revenue from the CRM. But let's go even more basic, because you can technically use target row as with arbitrarily assigned values for conversion. So all that means is, I'll use a simple example for home services, let's say your business decides that a repair call is worth X to you and an installation calls were to 2X whether that's true or not just disregard. But we say repairs worth x installation is worth 2X, you're basically arbitrarily setting a value to those two conversions that are different and you're communicating to Google that one is worth twice as much to me as a business as the other so that the AI bidding can go in and say we're going to be more aggressive for a higher value conversion and less aggressive for a lower value conversion. So technically, you can use this strategy and arbitrarily set those values. However, that's tricky, because you're kind of saying like, installations are always worth 2x. And repairs are always worth X. And it's rarely that cut and dry or that simple. So what we found are, we believe the best way to use a target rollout strategy is to actually integrate your CRM. So the system that you're booking all of your phone calls in and tracking all of your revenue in with your Google Ads system and have your revenue import into Google ad system. So the machine not only knows that it's a repair call or Installation Call, but this repair call was worth $1000. This installation call was worth $2000. This repair call was worth $3000. Right? Those don't necessarily fit the X or 2X example that I gave you, but the system knows, okay, these are of more value, these are of less value. And they can be more or less aggressively with really, really specific information about not just how many conversions do we want, but what types of conversions do we want? What quality of conversions or what value of conversions do we want. So the cons of this, I would say are, it's a little complicated to get it set up relative to some of the other strategies. It requires integrating systems. So usually Google ads in the CRM, or if you can do a direct integration between your CRM and Google ads, you can use what's called offline conversion import. So you basically have like an in between step where you have information, go from your CRM to a spreadsheet, from a spreadsheet to Google ads, you kind of have like that intermediate step. But again, it requires a bit more planning a bit more structure definition to what you're trying to accomplish. And so not maybe the best starting point, but I think a really good target like to evolve into with the machine learning strategies. And I think, ultimately, probably wherever ones heading over the next several years similar like E commerce, or E commerce has all this information for their conversions, not that just that you purchased. But how much did you purchase, what was the value of that purchase. And this is really the beginning of the evolution for home services into that kind of space that ecommerce has been in for a while. So that was really exciting. We could probably do a whole podcast on just that maybe we will.

Tim 16:31

This is really the future when Google can see that, you know, a home that has three bathrooms is more valuable to a plumber have that than one that has one bathroom and bid accordingly. That's really in some ways, what we're talking about here. And that's, that's the power of AI the power of what's coming in the future as Google gets in, as our integrations with Google and our CRM get smarter.

Ashley 16:56

Yep. And then the last few strategies that we touched on more specifically. So this is where I think more about like awareness, if awareness is your goal, this is probably the only times I would recommend these strategies. And there's certainly a time and place for that, potentially, but maximize clicks. So again, trying to get as many clicks as possible. However, when you target clicks, you're not necessarily targeting conversion, because clicks and conversions aren't the same thing. And no matter what anyone tries to say to you, or no matter what any pitch that you get, like about a marketing platform says clicks are only good if they result in business. If you get 5 million clicks and you make $0. From those 5 million clicks, they have no value to your to your business. Right. So the cons here, I think that the Pro is if you're going for awareness, this may get you like seen more often. But the cons are, Google's going to do whatever it takes to get clicks. And that means probably branching out into things that aren't quite as good a fit for your business and may not be as likely to result in conversions, which is where you're ultimately heading. So I'd be super careful about using strategy. I think that this it sounds really good people here and I'm like, Yes, that's what I want more clicks. But it's no, you really want clicks that lead to conversion. So you just got to be careful with that one, because Google's gonna really define it as more clicks equals better, not more quality, clicks equals better necessarily so. And then the last one target impression share same idea really about awareness. For people that have been in the paid search game for a while Google used to have a metric called average position, people really liked when their average position was 1.0, or very close with that meant you were first at the top of the page above everybody else. I feel like this is kind of where this is evolved to. So when you use the target impression share, you say Google show me at the top of page X percent of time, show me at the absolute top of page X percent of time, again, Pro is visibility, you're going to be in front of people often and that can be good for certain strategies. But position doesn't necessarily equal conversion. In fact, very high position often these equals spend a lot of money to be seen at the top of page all the time. And if that's an impression, that's not likely to turn into a conversion, that may not be the best use of your money. So again, just wanted to be super careful I will mention too, that we tested this about two years ago, this strategy a little bit to see how well it worked. And we were able to manually get to the number one on the top of the page more often. So we're at like 95% of the time when trying to be number one all the time versus the machine being set to a target impression share for absolute top of the page of 100% getting around 85. So even if you're trying to do this honestly that the manual process works better than the target impression share for the machine. So again, be super careful with these strategies, but there may be a time and place for this.

Elainey 19:50

The other thing with maximize clicks and target impression share is that you do not need conversion tracking to use those strategies. So if you do not have you your conversions set up to be imported back into Google ads, these are strategies that you can use, I think the biggest piece of advice is just really making sure that you're monitoring it, like make sure you, if you can't see your conversions in your Google Ads account, then make sure you're listening to those calls, and you're making sure the quality is there and that they're driving leads. There's a lot of other tools and buttons that Google provides that you can kind of use to assess quality. And you want to make sure that you're kind of using all those tools, even if you don't have the conversion tracking in place.

Tim 20:33

Alright, so we talked about kind of what each one of these strategies is, and talked about the pros and cons of them a little bit, let's get into some situations where you might want to decide to start using some of these strategies.

Elainey 20:54

I think one that we hear a lot is that I'm spending too much money, my marketing spend is too high, my cost per lead is too high. Nobody likes to spend money. So we hear that one a lot. And these are strategies that most of these strategies actually work really well in, in a limited budget situation. So if you're in a position where you want to reduce marketing spend, our number one advice is going to be let's try to maximize conversions. You know, let's try and get as many conversions as possible with whatever budget you would you want to set it as.

Tim 21:28

Can Google go find more for us?

Elainey 21:30

Correct, because if you just cut budget and you stay on manual, then you're gonna run into a situation where, you know, you're gonna bid that same amount of money on a person who's least likely to convert. So being on maximize conversions, and having a limited budget is going to force Google to only show your ad to people that are going to be likely to do business with you.

Ashley 21:57

As Elainey mentioned, with limitations, especially on budget, I think we've seen the biggest lift with some of these strategies when the business is super limited by budget when going from manual to a machine learning strategy. So because it does, the machine really does find efficiencies and kind of weed out, I'm gonna call it the garbage. I don't know if that's the best word but not as not as good stuff and go after the better quality stuff. And when there is a severe limitation, there is just so much to find those efficiencies. So if you are in a limited budget situation, like highly recommend getting on board with trying some of these strategies, and there's a lot of opportunity there.

Tim 22:34

Now, that makes sense, Ashley, right. But earlier, you cautioned, like if your budget is higher than the amount of traffic maximize conversions might end up spending a little bit more. But if you're at a limited budget, then it sounds like it really starts to gain the efficiencies.

Ashley 22:49

Yeah. 100%.

Tim 22:51

Cool. All right, what else.

Elainey 22:55

Let's say that you're, you know, you're in a really good spot with your business, you're really happy with your cost per lead, but you want to, you know, you're willing to spend more to get more. Situations like Target CPA are a really good situation for that. So we're able to hold our efficiencies that we've gotten, we're able to, you know, set a limit on that cost per lead, but we can add more budget into the account and see if there's more leads out there for that price point.

Tim 23:20

So here, you kind of know what the cost per lead is, because you've been running Max conversions for a while. And you're saying, let's listen, that's good. Can you go get me more of these? Is that kind of what's happening?

Elainey 23:35

And then sometimes there are situations where maybe you're unsure of what your cost per lead is. And you're just saying, I need more leads, and I'm willing to spend more money, you know, you may have a mixture of strategies in your account. It doesn't have to be set at the account level. So it is something that can be customized for different service category. And depending on how your paid ads account set up, I think the biggest thing with choosing a strategy that's going to work is making sure that you participate in the conversation with your marketing professional. So whoever is pushing the button in your ads account, and they're the ones that are managing your paid ads, talk to them about your goals, talk to them about what's going on in your business, share what you know, the more information you can provide with them, the better because there's so many tools and there's so many levers that we can pull that having that conversation is going to help, you know, help your marketing person reach your goals in the most efficient way possible.

Ashley 24:38

And like Elainey said, define your goals for your marketing agency or partner don't micromanage the strategy. So you need to tell them like where you're going, like how you generally want to get there, what's going on in your business, how much things cost, even in your business, what's a good cost per lead where you're profitable, like that kind of stuff is really, really helpful to share. I think where you can run into an issue is hopefully you found a marketing partner that you're working with because you trust that their knowledge on the marketing platforms and all these strategies is, is greater than yours but you as the business to find what you need. And then let them define the strategy to catch you there. If you micromanage like, no, no, no, don't do this, do that. And you don't really have experience with it, you can really shoot yourself in the foot as far as getting to the end results you want. And I think one of the worst things you could do as the business in this partnership is show up to a meeting with your marketing company with nothing to add to the conversation, right, like you want to meet to go through results you want to meet to talk about what you're going to next and you sit there quietly and add nothing to the conversation, let them share, which is fine, they should share with you. But if you're not having a two way conversation, I can tell you from experience with over 100 different clients, the best results that we can get for our clients are when we truly partner up together and share information between marketing and business back and forth. So just make sure like Elainey said, make sure you participate, but participate in a productive way, don't micromanage the strategy bring as much as you kind of the table about what your goals are as a business. And that's gonna allow your partner to do the best job for you.

Elainey 26:18

And even if it's a situation where you're like really happy with your lead volume, and you're not looking to grow, and like you're like, Okay, I'm in a space where like, I'm ready to stay like at this level for a while, still share that because maybe there's an opportunity to like, at least from our perspective, we love when we save our customers money, like if we're able to, like cut your marketing spend in any way like that's going to help you win, even if you're maintaining that lead volume. So like any piece of conversation, even if you feel that, like you're really content with things and your you don't want to change anything, there's still maybe opportunity out there.

Tim 26:54

Cool. All right, so last episode, we started to talk about a business that had really strong results by using machine learning and diving into some of these strategies. Can we can we go deeper into that and talk about the process to kind of roll this out and what happened?

Elainey 27:12

Sure. Um, so as a refresher, from our last episode, this business is located in a very large metro area, where cost per clicks are really expensive. And there's a lot of competition in Google ads. So there's a lot of it just is a it's it was a tough account just because of how expensive and how competitive people were. Because of the size of the market, there was budget limitations in the account also. So if we were to have an unlimited budget, I, it would be easily like millions of dollars in marketing spend. So you really had to find that balance. So I would say from start to beginning, this process took about a year. And in May or spring of 2020, we implemented conversion tracking. So we implemented the dynamic number insertion, we utilized CallRail and this allowed us to attribute down to the keyword level what terms were driving leads. So we were able to see in the account. Even just on manual bidding, where can we drive efficiencies? Are there any terms that we're advertising on, that are either really expensive with a really high cost per lead? Or if they're not converting at all? So is there any low hanging fruit in the account that we can eliminate the process of going through that and auditing that data does take a couple months? You know, there is seasonality in most home service industries. So you kind of want to work through that seasonality, where in the summer months, maybe, you know, things are a little bit cheaper. But then as we get into the fall in the winter, things are more expensive, and you kind of want to maybe cut more off the tree.

Tim 28:50

So I'm just gonna I'm just gonna simplify this a little bit, right? So we use this buzzword DNI and CallRail. So we are not an affiliate of CallRail or anything like that, but we use CallRail and and the reason we do is when we talk about teaching Google what a conversion looks like most of our home service business, a lot of our customers all of our customers, and we need to teach Google what a conversion is. And that means we need to set up conversion tracking, using what's called DNI. DNI allows new phone numbers to drop on the site. Every time a new visitor comes to the site so that Google can see who converted what keyword they typed in, what time of day it was, who this was, they already know the person and they this allows them to connect all the dots. We're kind of watching that for three months. We're not telling Google to go find us more, we integrated all that and then said, Okay, Google, here's a steady stream of information for you, right?

Elainey 30:02

Yes. Um, so, while still on manual bidding, like I said, we were able to drive a lot of efficiencies. But after meeting back with the business owner, they were still like the cost per lead wasn't profitable. We were still pretty limited by budget in Google ads. And at that point, we started rolling out the test for maximize conversions. It was a really long process of rolling this out, it took a few weeks to a few months to test slowly. So we did one campaign at a time, you know, our goal was really to get more leads for the customer and reduce the cost per lead. So we wanted to monitor this really closely to make sure that those goals are being met. Before we kind of did like a blanket change to the whole account. Like I said, this testing phase took about three months to roll out. But so at this point, now, if you're keeping track of timelines at all, we're about a year since when we first implemented CallRail, and being able to see that conversion tracking down to the keyword level. So looking at year over year, our conversions tripled. And our cost per conversion cut by two thirds. So to put some numbers to that, our conversions increased from 95 conversions in the month of May for 2020, all the way to 275 conversions the following year, and the cost per lead or cost per conversion on that dropped from $304 to $110. We did not touch budget during this timeframe. So Google did what it had to do on maximize conversions, we said, here's our budget, this is how much I'm willing to spend, get me as many leads as possible and Google really delivered. I think the important part of this is that we did spend many months auditing the account for efficiencies already on manual. So just something to keep in mind as you're working in your accounts. Automated bidding and machine learning strategies are really successful on good foundation. So make sure that your account is set up, your account has efficiencies built into it, and then your machine learning strategies are going to perform even better.

Tim 32:22

Okay, cool. So in this situation, here's one, we're on limited budget and maximize conversions was a complete game changer. It almost tripled conversions, while almost like reducing the cost by two thirds. So just unbelievable cost reducing cost per conversion. Yeah. So Okay. Let's take another example.

Ashley 32:46

Sure. So Elainey went through an example like she said, very large metro area, very expensive, very competitive, like densely populated with businesses and the in the industry. So this one's a smaller metro area, maybe small to medium sized metro was significantly different. There, there was competition, but it was very manageable. There were a few big players, but not a lot. It wasn't, it was not super, it wasn't the cheapest market or metro area, but also was by far not the most expensive when we looked at others in the industry. So kind of like a middle of the road situation. And we were working with a client that had been doing Google Ads pay per click for many years. So they were familiar with the platform when the average position metric existed, which has now been quite a few years since it existed, they became quite obsessed with being number one or at the very top of the page as often as possible. And they started to kind of attribute being number one, or at the very top of the page, like that was that was what meant what success meant I had to be number one, that's how I get conversions. If I'm not number one, I'm not gonna get the most leads I possibly can. It's not, it just can't happen. So this is what I'm obsessed with this is what I must do. And over time, he drove up his costs significantly, like overtime 10s of 1000s of dollars to give you an idea, so not like $50 more a month $100 more 10s of 1000s of dollars of increased marketing spend with very minimal impact to the volume of conversions he was receiving with the manual bidding strategy and the obsession with being at the very top of the page. So we worked hard with this client to convince them that they should try the machine learning strategies and see if we can bring costs back down while at the very least maintaining conversions possibly trying to increase conversions. And they did finally get on board. We got that going. However, they dictated with the maximize conversion strategy that they did not want to have any limitations on their budget because again, this idea of like being present all the time being seen all the time was really, really important to him and so he wants to know limitations on the budget. And what happened initially with the machine learning strategy is exactly what we don't want to see but costs went up even higher than they were on the manual bidding strategy and conversions stayed virtually flat. So it was just against spending more and more and more 1000s of dollars a month more to get the same number of conversions, which is obviously not the goal of any, any strategy. So we worked with him over time. And we're able to start backing that budget off a little bit and getting closer to like an ideal budget or a slightly limited budget, we're limiting our impression share just a little bit. And as we did that, we were able to increase the conversions and decrease the cost per conversion simultaneously. This happened about over a course of six months, this machine learning kind of evolution. And so to give you an idea when they first right before they got start with machine learning, they were at 240 conversions a month, and their cost per conversion was about $140. After we experimented with machine learning and started to pull back on that budget and not give maximize conversions, that unlimited budget, their conversions went up to 312 in a month, and their cost per conversion went down to 115. So they basically had a 30% increase, I'm sorry, yes, 30% increase in the number of conversions and about a 20% decrease in their cost per conversion going through this process. And I will mention, this happened between the month of June and December. June being that early portion with fewer conversions December in this industry is actually one of the slower months of the year for conversion. So not only did they grow their conversions as they went from June to December, but they hit a high by a lot in December for a month that's typically far slower than June, like on a seasonal basis. So really significant change to their volume of conversions as well as their their cost per conversion, not quite as significant as Elainey's example after really a shining star example, but still huge impacts to their business and their bottom line.

Tim 36:51

So here we have a mid market, it's still a big impact on their business. So we had two instances of maximize conversions that work really, really well. How about TCPA?

Elainey 37:08

Yeah, so for Target CPA, we found that it works really well for reducing cost per conversion when that metric gets too high, or if we're wanting to pull back at all. So we have two different situations. One, this is a very small market were randomly out of nowhere for no reason that we can really determine cost per click started going up substantially, which also costs the cost per conversion substantially. And we're talking like from a $10 cost per click to like a $40 cost per click like it was in within a week's time. There was also no new competitors really in the market. So we couldn't tell like if it kind of would make sense if there's new competitors, and they're getting bidding more aggressively than like, our bids are gonna go up. But we didn't really see any of that happening. So we put a target CPA in place, they already were on maximize conversions. So we layered a target CPA in place and we set their cost per action or cost per conversion at where their previous average cost per conversion was before things got really volatile. And almost instantaneously, things balanced back out. We saw conversions, come back to a normal level. And we saw cost per conversion come down.

Tim 38:29

So that was a brilliant strategy, by the way.

Elainey 38:33

Yeah, that really was like, I think a really good example of that leash, where like Google was just running wild, it was doing whatever it wanted. And we needed to kind of pull that back in.

Ashley 38:43

Elainey, did you see any impact to the volume of conversions in that example, when you were able to bring it back down into the the old kind of cost per conversion?

Elainey 38:51

Definitely, I think at first, the target CPA that was set was probably maybe a little bit lower than it needed to be. We did see conversions drop a little bit, but this was a really small market. So it's really hard to tell the difference between week to week on a conversion volume when the data set is so small. After a few months, we changed our TCPA or set cost per action to be a little bit higher, and we saw conversions come back. So you know, we still have that TCPA in place because we don't necessarily trust Google because of how their behavior was, but we will give it a little bit longer of a leash to try and get more leads. Another market which this one is probably a more common situation where it the cost per conversion was creeping up. Just like Ashley said earlier, it's normal for cost to increase in any market year over year. But the owner of this business didn't feel comfortable putting more marketing budget into the account. So we utilized a TCPA to force Google to go after cheaper conversions. So We set our target CPA to be a little bit lower than what the average is to try and squeeze more conversions out without having to add more budget to the account. So if you kind of looked at a trendline, the spend in the account would have stayed flat, but our conversions went up just a little bit. But you know, all every little bit helps. And looking at things year, over year, we do see some pretty strong results. So TCPA is a little like you, you kind of see the results better when you look at them, like more long term, or maximize conversions are definitely where you see that more substantial change.

Ashley 40:39

I think with any of these two, like there are certain accounts where you feel like one strategy is going to work really well, or maybe another and you go and test it. And a lot of times it works out how you expect, but sometimes it doesn't. And so it's always really important, like test and evaluate, test and evaluate if it's not working, like relook at it, come up with a new strategy, and then test and evaluate because it's not a one size fits all. And it's not always like something that worked in one place that looks very similar to another, they don't always work the same. So just know that like, you have to be open to testing and evaluating those results. And sometimes the first try isn't the best fit, sometimes it is the best fit. And then over time to sometimes a strategy that works really well for 6, 9, 12, 18 months, all of a sudden stops working as well. And you have to go back to the test and evaluate kind of, you know, mindset. So you got to just always be open and willing to either stick with the strategy or switch strategies if it's not working stuff.

Tim 41:40

Let's just talk about the order of this a little bit. I know in most of our accounts, we started them with manual bidding, then we would go to maximize conversions. And then when when situations arose, we would use TCPA. Right? Well, when we do new accounts, so we typically launch with manual bidding, or are we using a we experimenting with the machine learning like almost right away now.

Elainey 42:06

I think you definitely start with manual bidding. Because if you're just start with maximize conversions, like it's hard to tell, like if your results are good or not like you don't have anything to really compare it against. So we definitely depending on the size of the market, like anywhere between like one to three months worth of data on manual strategy. And if we see certain indicators, like maybe you're limited by budget, super early on in your ad account, then you may lean more towards using a machine learning strategy, like maximize conversions earlier on. But typically, we need a little bit of data so that we can tell if it's, you know, if it's performing better or not.

Tim 42:48

And those are the strategies we use in the home service industry most frequently right, we're more about maximize conversions, and cost per action than we are about some of the visibility strategies right?

Ashley 43:04

With the conversions, a lot of the conversions that we prioritize are lower funnel conversions are closer to actual purchase conversion. So we're a lot of the spend is going not towards the awareness stage or like that earlier funnel, but the most bottom of funnel term. So yeah, we're typically targeting a conversion as opposed to like an awareness kind of strategy. Also, like, the stuff that we've experimented with in Google, when we tell them, we want a lot of clicks, they get a lot of clicks, but we're not really happy with the place that we're getting. So we haven't had tremendous success, or felt that the quality of the results are really up to where we want them to be with some of those other strategies. So we've kind of moved away from them, for the most part.

Tim 43:45

All right, any final words?

Elainey 43:52

I really think it's be willing to test new things as Google rolls them out, like be willing to try new things. And if it doesn't work, maybe come back in a few years and try again, because that's definitely our story. You know, we tried many years ago, wasn't thrilled. But now here we are two episodes deep, really excited about it. So I think it's um, you definitely give it a try.

Ashley 44:16

I was just going to say, we alluded to this last time, but there may be a point in time where manual bidding isn't an option anymore. So you know, experimenting with some of these machine learning strategies while you have complete flexibility over which strategy you pick, even if you pick one campaign to test it in just to get your feet wet and get an understanding of it, I think is a really wise strategy. This is being pushed more and more and utilize these strategies more and more by more businesses. So you just don't want to be left behind or like in the dark about how they work and how they can help your business.

Tim 44:53

Awesome. So you know, folks in the home service industry We all know how how expensive. You know, Google has been getting. How much clicks up and rising through the pandemic, as more business became available. A lot Google's cost rose a lot through that. And the education that we got here from these two ladies who really know this is absolutely invaluable. So thank you very much. That was just an awesome education on the automated bidding strategies that are available from Google. So thank you. And thank you audience for tuning in. And we'll see you next time.

Episode 7 Cover Art
A2O Digital
Office: (973) 283-3111
220 West. Parkway
Suite 9

Pompton Plains, NJ 07444