Google Ads remains one of the most popular ways to advertise online. In 2019, advertisers spent just under $135 billion on the platform. One of the most important decisions you’ll have to make in properly managing your Google Ads campaigns is what your primary objective is, and how to choose your bid strategy accordingly.
Today, we will dive into the various bidding strategies that are available to you as of this moment, and when it’s best to use each. For the purpose of this post, we’re assuming that you are using a standard Google Ads account with full access, and options, and not an Ads Express account.
Automated Bid Strategies
In recent years, Google has tested and released many machine learning bid strategies, designed to help enhance your results based on common objectives advertisers have for their Ads campaigns. As such, there’s often debate among paid search managers of if these are worthwhile, or if it’s just something else Google is pushing on advertisers to further line their own pockets.
First, we’ll get into each of the available bid strategies, and why you may want to use it, then we’ll get into the pitfalls, and general recommendations.
Target CPA
Target Cost Per Acquisition (CPA) – this bid strategy is intended to pay attention to the actions you define as a conversion within your account, and try to make it so that you get as many conversions as possible for the lowest CPA possible. This works great when you have a clearly defined group of conversions that have enough volume to allow the algorithm to pick up patterns, and be able to refine itself.
How many conversions do you need for Target CPA to work?
Officially, Target CPA can be implemented right away, out the gate, without any historical data. Most (if not all) paid search managers would advise against running this strategy immediately on a new account without any data. Without a baseline present, the algorithm will be all over the place trying to figure out what to do, and you’ll likely spend way too much on certain searches, and blow your budget.
Recommendations for building your data history tend to range between 15 – 50 conversions within 30 days. Our ever political answer is, “it depends.” If your conversion set is well defined, and restricted to those actions that mean the most to your business, and not necessarily full of high-volume behaviors, I would err on the side of 30+ conversions per month, and have at least 2 months of that kind of consistent performance under your belt. I would consider certain activities, such as form fills, and tracked phone calls as this type of conversion tracking.
If you are considering certain page views, dwell time, or other types of behaviors that can lead to higher volumes of conversions, the lower quantity of 15 within 30 days is sufficient, as those types of conversions are easier to obtain, and will naturally refine more quickly.
Target ROAS
Target Return on Ad Spend (ROAS) is an automated bid strategy that is intended to be especially useful for those in eCommerce. It requires that conversion values are provided in Ads, so that the algorithm can pick-up the action value for each conversion. For eCommerce sites, a common example would be a transaction value of a completed purchase that resulted from someone clicking on your ad, then going forward with a purchase.
Target ROAS bidding is best when you have a way to directly attribute a dollar amount in sales to the paid search click. If you are able to track offline purchases back to the click through lead submission tracking, be sure to upload that data in regular intervals. If your site allows for direct purchases, that is the best case scenario because there is no question that you are able to track those purchases automatically with conversion tracking provided by Google Ads.
When should I use Target ROAS?
First, ensure you have conversion tracking set. Second, ensure your conversion values are set properly, or are being sent in from your site based on a tested dynamic variable that will correctly import the value into Ads for each transaction. The conversion values should be based on actual dollar value to your business. If it’s left as 1 for everything, this is essentially the same as Target CPA.
Many guides (including that provided by Google for certification testing) suggests that as long as ROAS is positive, keep spending, and you’ll be in good shape. An important caveat that seems to get lost in that is determining how ROAS is calculated. If you are taking the total value of a completed eCommerce transaction, and considering that as the amount for calculating against ROAS, I would argue that at 100%, you lost money. That means, if the transaction was $30, you spent $30 to get that sale. You lost money because in this scenario, cost of goods sold, shipping, etc. are not considered. Even though this is considered a “break even” ROAS, your business is behind because all of the revenue earned is only covering the advertising cost.
There are a couple of ways to account for your costs for ROAS. First, is to set your target based on including a rough estimate of your gross profit amount. For example, if your average margin is 40%, your break even point would be 250% ROAS. The second way is more accurate, but can be more complicated (depending on your eCommerce platform) – adjust the value that is sent to Google Ads so that it already accounts for the associated costs, so that it will account for fluctuations in margin if your products have large range. This would make it so that ROAS of 100% is even, and accounts for various profit margins.
Maximize Clicks
As the name would suggest, this automated bid strategy tries to get as much traffic to your site as possible with the given budget. This strategy will chase as many clicks as possible with your budget, no matter how aggressive or inexpensive those clicks may be. The upside is that you are able to set a CPC limit, so you can make sure the max bids are kept in check.
When should I use Maximize Clicks?
If you are looking for getting volume of visitors to your site for the purpose of building a remarketing list, or if your site naturally converts well, and you simply need more people to get to it. The quality of traffic for these campaigns greatly depends on your keyword refinement, and ensuring the most relevance possible. If you have too broad of a keyword list, you will find the traffic to be somewhat, “junky” based on user behavior metrics, and that there is a large volume of people finding you with search terms that aren’t as relevant.
Use maximize clicks if you do not have a defined conversion set that generates enough data volume to make efficient use of Target CPA. Also, maximize clicks can be helpful for new accounts that needs data to learn what works, and what doesn’t – you can set CPC caps to ensure the algorithm doesn’t go crazy, and you’ll get volume much quicker than with some of the other bidding strategies.
Maximize Conversions
This is a bid strategy that is similar to Target CPA in the sense that it’s trying to find conversions, but without the restraint of trying to get conversions at a particular cost.
When should I use Maximize Conversions?
We typically use Maximize Conversions as a precursor to Target CPA. If a campaign does have some historical data to work with, but doesn’t necessarily have high volume of conversions, this could be a good method in which to get the campaign further optimized toward getting more conversions without the constraint of having to meet a CPA threshold.
This strategy can lead to seeing a greatly increased CPC metric. The payoff is where you will see more conversions as a result, and be able to get enough conversion volume to be able to switch over to a target CPA model, where you can have cost controls in place toward getting more conversions at a lower cost.
Maximize Conversion Value
This strategy is much like Maximize Conversions, but useful if you have varied conversion values. If certain conversions are worth more to you than others, be sure to have that reflected in the conversion value in your tracking settings, and use this strategy instead of Maximize Conversions.
This is a good strategy as a precursor to Target ROAS for eCommerce, and for lead generation where you may have a different value for certain kinds of leads over others. For example, if you find more value in a consultation form fill compared to a general contact form, have the conversion value reflect that so that the algorithm will also value those actions according to what you see as more important for your business.
Target Impression Share
If you want to show up in a certain percentage of searches, and bid according to that, only, this is for you. Depending on the type of campaign, and how much competition there is, your impression share naturally varies.
In all honesty, this is probably our least favorite automated bid strategy, and we rarely use it. We prefer to use strategies that employ more performance-based objectives, such as those that directly lead to sales or leads. Target impression share does ensure you show up if people are searching for related keywords, but with nothing else particularly factoring in. We would rather focus on being more aggressive in searches for those that are most likely to convert, so you’re always in one of the top two positions for those, rather than potentially having a lower overall position for more searches.
Additionally, it is quite easy to pay attention to what your impression share is, and adjust your other targeting accordingly to make sure you hit any benchmarks you may have for it. If you find that your share is too low, you may need to adjust your budget, geographic area, or keywords so that you can improve the impression share. Using this bid strategy would act as more of a bandage to mask other parts of the campaign that may not necessarily be optimized. If you can hit your impression share marks while using most any of the other strategies that lead to machine learning toward conversions, that will better serve your needs.
Pitfalls of Automated Bidding
Automated bid strategies are not perfect, by any means. In certain cases, it does the job well for a particular purpose, but falls short in other aspects that may not be deemed as important as the primary objective. Some common issues:
Inflated CPC
With several of the bidding strategies, average CPC can hit levels that you would never dream of, had you set a limit for yourself. We have seen, in some cases, bids hitting 3-6 times as high as what you would normally pay as an average in a manual bidding scenario. This is often a result of the AI being aggressive for certain clicks that are more likely to lead to the result you are looking for, or early on in the learning process, where the machine learning is still trying to figure out what works best for your account. Either way, some of the CPCs you will see with an automated bid strategy can be unnerving, and you’ll have to have a strong constitution to look past it if the strategy is, in fact, working well for the greater good.
Learning Time
Like any machine learning, it takes time for the system to properly optimize your account for the objective you are seeking. Before you even think about using most automated bid strategies, you need to establish some historical data to work with, and then you’ll need to endure a week or two of what I would describe as, “wonky performance,” in the name of refining the bid strategy specifically for your campaign. During this learning period, you can see some days with considerably higher CPCs with lower clicks, other days with lower CPCs and higher volume of clicks, costs all over the place, and enough other things that would drive someone with OCD tendencies crazy. If you are able to manage to see the light at the end of the tunnel after the learning time, you should see performance better than what you were seeing before implementation of the automated strategy, in relation to your objectives.
Decreased Performance
Sometimes, the automated bid strategy simply does not optimize, and you will find that you’re just spending more money for clicks that aren’t converting any better than before. This is usually due to not having enough data to optimize. If you try to switch to an automated bid strategy because you inherently think automated = simple, you will be in for a rough reality if you don’t have frequent enough usable pieces of data to help the machine learning advance.
Manual Bidding
Manual bidding is exactly as you’d expect – you set the bid caps, and that’s that. Google also has Enhanced CPC, which is a form of manual bidding combined with some basic machine learning to optimize for either conversions, or conversion value. Either does allow you to set your own bids, but Enhanced CPC does take some liberties with your bid caps in order to chase conversions.
When should I use Manual Bidding?
Many PPC managers like to start off a new account with manual bidding to ensure the account is setup in a way that doesn’t have shocking CPC metrics, and to have as much control as possible to setup the account for success.
That being said, manual bidding needs attention, and regular testing. With the constraints placed on manual bidding, you can easily lose out on conversions if your bid is set too low, and you could blow your budget quickly if your bids are set too high. These are areas where automated bidding helps, if the data exists.
Use manual bidding when you are setting up a new account, or campaign so that you can acquire the data needed to move onto an appropriate automated bidding strategy.
Final Thoughts
Automated bidding strategies with Google have improved dramatically over the last few years. Even still, many PPC managers still caution against the use of them, likely because of a past bad experience with it, or simply for the fact that it does hand some control over to Google.
We do endorse the use of automated bid strategies with Google because of its inherent ability to utilize more data than you could possibly see on your own, and that there are several different strategies available that can work toward the benefit of many different objectives.
When you do make the switch to an automated strategy, you will likely have to endure some moments of shock and doubt that it will work properly because of some of the immediate differences you see that may boggle the mind.
When we made the switch of several campaigns from Enhanced CPC to Target CPA for lead generation purposes, the initial results were disastrous. After 3 days, the thought of, “we made a huge mistake, and if this keeps up, we would fire ourselves for how this is doing,” was commonplace. Then, it started to even out to the same performance as before. After 2 weeks, we saw that the campaigns were starting to convert more, and CPA metrics started to reduce by over 30%. CPC metrics were still slightly elevated compared to before, but we were happy to see the additional leads come in at a lower CPA, as that is the primary objective for these campaigns.
Several months later, average CPCs are now lower by 20% year over year when we were running with Enhanced CPC, and CPA has maintained the same levels as initially seen, where it is reduced by over 30%, on average. We do still see individual bids that are obnoxious, on occasion, but the overall results continue to speak for themselves, and our clients have continued to see greater success through their paid search efforts.