Facebook has quietly published a Help Centre post stating that the Campaign Budget optimisation feature will become the default budget setting in September 2019 and will not be able to be turned off.
Here is the full post from Facebook detailing what we should expect from this change.
What is changing?
Campaign budget optimisation was rolled out over a few months in early 2018 as an option to apply one budget across a whole campaign and allow the platform to optimise spend per ad set automatically.
Similar to ad distribution optimisation, which is now the default setting with budgets controlled entirely at ad set level and optimised automatically across ads, this feature gives the option to allow the platform to learn where best to spend budget within the audiences you have built in your ad sets. This should theoretically lead to more efficient ad spend, as the aim of automated optimisation is to drive as many actions as possible with costs which align with your bid strategy.
Currently, the feature is optional: you can choose to leave it off and set your budgets at ad level, and then set a campaign spending limit if you wish. Or, you can apply campaign-level daily or lifetime budget and set minimum and maximum spends per ad set to retain a level of control.
However, it seems that when campaign budget optimisation is rolled out as the default setting in September there will not be a minimum/maximum ad set spend option as there is no indication of this in Facebook’s post, nor will you be able to opt out of setting budgets at campaign level.
Budget-split consideration is an integral part of deciding how to structure a campaign for efficiency and to allow testing, so this change will shake up the campaign-building process considerably and advertisers will need to reassess how they approach this.
Benefits of switching to campaign-level budgets
#1: More efficient spend
The aim of automated optimisation is to expand reach and allow the platform to serve ads to those most likely to convert – and cheaply – first from a larger pool of users, meaning it will be able to keep costs down for longer before moving on to the more expensive conversions. By applying one budget across all ad sets, the platform can optimise campaigns automatically across the entire pool of users you are targeting, rather than just by ads per audience. This allows you to reach those people most likely to be cheapest with no restriction on spend per audience.
#2: Automated changes to budget in response to performance changes
Currently, when budget is set at ad level this needs to be adjusted manually if an ad set starts to underperform and become more expensive, meaning that performance depends in part on an advertiser’s ability to make these adjustments asap. We all know how busy the life of a social media advertiser can be and there may sometimes be a delay before performance changes can be spotted and adjustments can be made.
Having budget set at campaign level will mean that the platform will automatically adjust its spending behaviour to account for changes in ad set performance and keep costs consistent.
#3: Automatically adjusts spends to audience sizes
It’s difficult to know exactly how much budget to apply to an audience to ensure there is a balance between good performance and ad frequency. Having budget set at campaign level will mean that this will be automatically adjusted rather than an advertiser needing to test budget based on the estimated audience size and ad set reach.
#4: Simplified campaign build
With the budget-splits by audience all being automated, the need to calculate separate budgets for each audience and make sure the total is always correct whenever budgets are amended will be removed, as there will only be one value to amend, making ad set creation more streamlined.
When might CBO cause problems?
It’s becoming standard practice to group elements of a campaign together to allow automatic optimisation. The platform can adjust spend per audience (or placement, device, etc.) based on performance in real-time and not waste budget where it isn’t being used efficiently. Options like automated placements are increasingly being used as Facebook’s targeting algorithm continues to improve.
However, most Facebook advertisers still appreciate the freedom to split budget across audiences when they see the need to. There are a few instances where problems may begin to arise when this freedom is taken away:
#1: If the audiences being tested have different sizes and values
So far when testing campaign budget optimisation, we have seen that the budget split across audiences depends in part on the audience sizes. This makes sense as a small audience will have a limit to how much it can spend before ad frequency gets too high and costs increase, but for acquisition in particular this may become a problem if a smaller audience being tested has a higher value in terms of driving meaningful conversions.
For example, what about if we want to test an ad set of 1% lookalikes of past customers, returning customers, etc. against an interest-based audience for an e-commerce campaign? We can infer that the lookalike is likely to be more valuable in terms of both conversion rate and lifetime value. However, the interest audience will likely be much bigger so Facebook will start to favour this ad set as it will drive a higher volume of actions.
The ad sets cannot be combined into one to get around this as in the case of different audience types this would make the total pool of users smaller rather than larger, i.e. only people who fall into both the lookalike and the interest audience would be included. To get around this issue if it arises, the audiences would need to be split into separate campaigns and allocated separate budgets. Hopefully Facebook will work to refine its algorithm further when campaign budget optimisation rolls out to avoid the need to do this in the future, making CBO more sophisticated and able to create a balance between value and volume.
#2: More complex campaigns where specific budgets need to be applied to certain audiences
Some advertisers may require specific budgets for specific locations, or maybe age groups or genders if testing for conversion volume vs lifetime value. In this case, separate campaigns will need to be made each with one ad set, meaning ad accounts will end up being bulkier with several campaigns now needed per objective where previously there was only one.
#3: If the campaign objective is set to a specific action, which may not align with engagement or click-through
Facebook’s targeting algorithm has vastly improved over the last few years and is continuing to do so, but it is clear that even with objectives set to a specific on-site action such as a purchase, ads which drive a high level of engagement or clicks are often still favoured when it comes to taking ad spend. Too often we see ads which have received more initial comments or clicks starting a snowball effect where they start to take most or all of the spend without the other ads ever getting a chance, which can lead to campaigns driving less of the more meaningful and valuable actions as users who are engaged or likely to click are not necessarily the most likely to convert or have a high value.
Being unable to set budgets at ad set level may mean that we begin to see a similar effect in our ad sets, with more budget going to audiences that are better performing in terms of engagement and driving traffic but are less valuable in terms of driving conversions. Facebook will need to improve its algorithm for conversion optimisation further if we are to trust it to drive the actions we want with our budgets.
This will be a controversial change with users unhappy that their ad set spend control is being removed. However, as with ad distribution optimisation, it is expected that advertisers will eventually adapt their campaign-building strategies and campaign-level budgets will become the norm. In the meantime, there are likely to be times when we may not agree with the platform’s optimisations and feel the need to split campaigns. Facebook will need to improve campaign budget optimisation further for it to be trusted, and we have eight months to determine how to use it efficiently to get the best out of our campaigns. If this is pulled off, then we may look forward to more consistent and rewarding campaign performance in the future.
You can read Facebook's post here.
