There are two main things you need to be able to achieve when managing a large scale Google Ads account:
- Effective management
- Incremental improvements
This post aims to deal with the former, and its two principal components:
- Meaningful reporting
- Quick and appropriate changes in response to performance
Underpinning these, are two things: best practice and technology.
Below I’ve outlined the industry best practice and the fundamental tech you need for the core phases of Google Ads account management: building, restructuring, monitoring, optimising.
By following these guidelines you ought to be able to smash the management of your large scale account, thereby freeing up time to focus on number 2; planning and implementing changes to allow incremental improvements.
Wherever possible, keep all your campaigns in one Google Ads account. This will make reporting and tech significantly easier to manage.
Any dimension that you might be tempted to split by account (for example, location, product, or audience) you will be able to differentiate at campaign level.
Occasionally separate accounts will be your only choice – for example, if you have to have separate billing entities for different regions. Alternatively you might have multiple agents working on your account and you do not want to give them all edit access to everything.
In these cases, try to create as few accounts as possible.
NB – there is a limit of 10,000 campaigns per account. There really is no reason why you should have that many campaigns in your account so this shouldn’t be a reason for having to create multiple accounts.
Your campaign structure should reflect two things:
1. The different parts of your business or product offering which are meaningful for you to report on.
For example, if you are an ecommerce business selling men’s formal wear, you are likely to want to know the performance of all the keywords relating to your suits, separately from those for your cufflinks. That would warrant separate campaigns.
2. Things you expect to perform different to each other or you have different targets for.
If you sell different brands of suits for example, but don’t have different performance targets for different brands, then there is no point in having a separate campaign for them.
Difference in performance also extends beyond product offering.
For example, Brand keywords will perform very differently to Non-brand. That’s because people searching for your brand have a different intent to people searching more generically, and your brand CPCs will be much lower.
We would always advocate for splitting Brand and Non-brand at a campaign level.
The same principle follows for Competitor campaigns when you bid on someone else’s Brand terms. Again, the intent of these searches is different to Non-brand, and your CPCs will be much higher – this means you need separate campaigns for these keywords.
As an aside, I’ve written another article here about Competitor bidding if of interest.
The other crucial split you need to see at campaign level is match type. You will see different performance with different match types so they should be separated at a campaign level.
Exact match for example usually has a lower CPC due to a higher quality score with more relevant ads; BMM the opposite.
Further, the best practice is to be spending approximately 70:30 of your budget on Exact:BMM keywords – and you control overall budgets at a campaign level.
You should aim for single keyword ad groups. This means that all the ads within an ad group are triggered by only one keyword – which then means that you can make the ads very specific to the search coming in.
Not only will that increase your CTR, but over time will increase your Quality Score, which can then knock down your CPCs.
If you want to build out single keyword ad groups, having a Campaign Builder tool to do it for you will save you a lot of time and hassle.
Most tech driven agencies, ourselves included, will have an in-house tool that allows you to decide on your keywords and dynamically insert those into your ads.
It will then generate all the sheets you need to upload your new single keyword ad group campaign into Google Ads Editor.
If you don’t have an agency with this capacity, or you are working in-house, the best proxy would be to create a spreadsheet with formulae in it to form the fields you need for your upload.
There will undoubtedly be a lot of manual work involved in setting this up however, and QAing the outcome.
This is very simple: you should be checking performance and spend every day.
I’ve seen accounts which spend tens of thousands of pounds a month blow through their budget in a week due to not being checked appropriately.
Even if you are just confirming what you already know, when dealing with large scale accounts that can have serious volume running through them, I can’t stress enough how important it is to check budgets each day.
You should also check performance every day.
Between the interface and Google Ads Editor, when you are dealing with complex accounts, it’s very easy to accidentally pause activity that you didn’t mean to.
It’s also very easy to not notice that until you see in a weekly report that volume has dropped off a cliff. That’s why it’s essential to make sure you are keeping on top of any fluctuations in your account on a daily basis.
Be particularly diligent when launching new activity.
It’s very easy, especially if you are not using rules in Google Ads Editor, to accidentally set activity to target worldwide for example, or set a budget that’s too high.
If launching new activity in a large scale account, you ought to check it the first hour it goes live, two hours after that, and then each day for the next 3 or so after that.
There are a few key pieces that if you set up will make your daily checks so quick that there really is no excuse to not do it.
Perhaps the single most important thing to be able to quickly and efficiently monitor is your spend. Sometimes performance won’t be the best (fluctuations are part and parcel of PPC life), but there really is no coming back from an overspent budget.
Again, most tech-focused agencies should be using a budget tracker that can monitor spend at a platform, account, campaign, and label level.
Any decent tracker should allow you predict pace of spend, show actual vs predicted spend, and projected spend given your average daily cost over X days.
Unfortunately most of this tech is usually kept under-wraps, but our budget tracker is about to become public – watch this space!
Once you have a budget tracker in place you can check it quickly and easily every day, and the chances of your spend going awry are so much lower.
When dealing with a large and complex account, it might be hard to spot anomalies in performance that can be indicative of something wrong, until it’s too late.
That can mean spikes in CPA, drops in impressions, or any out of the ordinary fluctuation in any metric that you’re concerned with.
The key to catching issues before they escalate, is having an anomaly checker set up. There are a few different open-source versions for this that you can find in my colleague Wes’ super helpful post: 120 Adwords Scripts to Supercharge your PPC Campaigns.
Check out the options between 110 and 120 to decide which checker is the most appropriate and useful for you.
You ought to be able to set up a checker that detects changes in metrics at different levels of fluctuation for different campaigns.
It will take a week or two to figure out how to set appropriate boundaries for these tools. For example, a 40% drop in clicks between any two days may actually just reflect day-to-day fluctuations and not be a cause for concern.
It is definitely worth the set up time. Using an anomaly checker can be the difference between noticing your Brand campaigns were turned off for a day, or for a week, or more.
If you are running an account with multiple product lines and/or landing pages, if one starts throwing a 404, who knows when you’ll catch that.
Maybe that landing page leads to one of the main revenue drivers for your account, and you notice a sharp drop in conversions day on day during your daily performance checks.
Maybe it’s for a less important product, and you don’t notice until you come to your QBR and notice that YoY revenue has dropped by nearly 100%.
A URL checker is so helpful for this. There again is an example in Wes’ article, 120 Adwords Scripts to Supercharge your PPC Campaigns, at number 112.
The advantage of this URL checker is that it can also be set to report if a landing page is showing any specific text – “out of stock” for example. In this case, you can quickly divert budget away from a product you can’t sell, and also alert your client or ecommerce team.
A reporting solution
This warrants a whole post in itself, so I won’t go into too much detail. Essentially, if you are running a large scale PPC account, or expect the account to scale, you need a more advanced system in place than pulling reports from the interface.
Tech-based agencies will either have an in-house system that works via the Google Ads API, or they will subscribe to third party software. Funnel, Data Studio, and Report Garden are among the top contenders for that.
Being able to show collated results across channels, campaigns, labels, you name it, is essential to spotting any issues in your results quickly. It will also make your life so much easier if you are checking performance every day.
Make sure you choose a solution that can pull in data from all the platforms you need, and ideally that can generate automated reports. Often these can be sent directly to your inbox, which again can be a significant time saver in the long run.
Base all your decisions on data.
That might sound obvious but we all get carried away with hunches some times. Even if you think one ad copy sounded much more appealing than another, if the other version has a better conversion rate, you need to respect that.
Also bear in mind performance ceilings – boosting the budget on a campaign that isn’t limited by budget is unlikely to produce any better results. Nor is increasing your CPCs for a strong keyword if it’s already at 100% top IS.
If you are running a large account, then you should have plenty of data to work with. Don’t forget though that there may be some valuable insights from results in other channels that can enhance what you are doing.
I’ve written an article here with some suggestions on what learnings you can apply cross-channel.
This really goes back to basing your decisions on data. Don’t be tempted to make sweeping changes within an account without testing them first.
Ideally that would be in an AB format – there’s an article we’ve written here about how you would go about doing that if you want to check it out.
When managing a large account, the key thing will be to make sure that you are checking regularly for statistical significance. The more data you have the quicker you ought to achieve an outcome that favours one variant over another.
If you leave your test running for too long, you’re risking the performance of your account on a weaker variant. You’re also losing opportunities to improve performance through new tests.
Use bid strategies.
Test them, at least. You might have some natural resistance towards switching to bid strategies, but large scale accounts are perfect for testing them in.
You ought to have enough data for Google’s machine learning algorithms to make sound judgements on when to bid up or not. We’ve seen great results in generating more volume at the same efficiency from using tCPA and tROAS in particular on large accounts.
It will also free your time up from optimising at a keyword level, which is neither sustainable nor wise for large accounts.
Quality score tracker
If you are managing a complex account, optimising bids on tens of thousands of keywords, and moving significant budgets around, you want to be able to tell whether you are making a long term impact.
You also want to know if the types of changes you’ve been making have been a success, so you can apply that knowledge to other accounts or other parts of the same account.
A good quality score tracker is essential for this, as there is no way to see historic quality score within your Google Ads account. I like to think of it has a health check on the account, that you can track on a day-by-day basis.
Fortunately, we’ve just made ours open source, so feel free to download the script from Search Engine Land.
The script tracks your quality score over time within a Google Sheet, including individual components of quality score (expected CTR, ad relevance, and landing page experience).
This means you can easily focus in on exactly what improvements you have made, or need to make, on your large scale account.
Again this is covered in some detail in my post on AB testing, but to reiterate: if you want to split traffic as fairly as possible between any two elements in your account, the most reliable way to do this is via an AB script.
Wes mentions one script in particular at number 34 in his 120 script article which runs using labels. It alternates between the two variants, so over a 2 week period each variant has run for each hour of every day.
There are alternatives to this in the interface, like using “rotate indefinitely” for ads, but it’s not great at actually splitting traffic. There is also the drafts and experiments feature, but using that relinquishes some control.
If you are running a large account, you’ll likely have a lot of tests running and historical results also. The ideal solution to this is a dashboard where you can quickly see the results from your experiments – both on-going and historic.
We have an in-house tool that does this. It also has the functionality of looking at different variants’ performance during different time periods so you can ascertain if any changes you’ve made have had an impact.
Most usefully, it calculates once you’ve reached statistical significance for CvR and CTR.
If you don’t have access to tech like this, you could set up your reporting software to compare different labels over time periods. You can also use a site like this AB testing calculator to determine once you’ve reached statistical significance.
I have never worked on a large scale client where the conversion rate remains uniform across different hours of the day or days of the week.
What’s more, if with a large client you are seeing time of day and day of week fluctuations the chances are they are meaningful, given the amount of data you are dealing with.
The Google Ads interface has limited capability when it comes to changing your bids hour by hour for each day. A better way of handling it is by using a 24/7 bidding script.
What’s even better than that, is using another sheet to calculate the bids for you.