Introducing Machine Learning to Your Paid Campaigns.
Paid Search, Paid Social, Performance Marketing, Ecommerce Strategy
Machine Learning and AI is a huge topic in marketing and is set to still be a huge talking point throughout 2023 - but it can be heard to break through all the noise and opinions in order to find out exactly how it can benefit your brand.
Both Google and Meta have been developing their machine learning capabilities over the past few years with updates to campaign types, build and management. If you are looking to get started with Machine Learning, this article will take you through the basics of what is all means and how it can impact performance of your brand's digital campaigns.
If you want to make the best use of Google's machine learning and AI capabilities, Performance Max campaigns are going to be the best place for you to start. In Google's Marketing Live 2023, more AI features and updates were announced but we believe you have to get the basics right first before you can start testing some of the more interesting AI reliant techniques.
What is Performance Max?
Performance Max is a goal based campaign type where advertisers are able to access the entire Google Ads inventory within a single campaign. It's not designed to replace your current search campaigns but rather complement those campaigns. A single performance max campaign can serve ad placements across Google's whole suite of channels; including Youtube, Display, Shopping etc; to help broaden your audience, take more advertising real estate and find more consumers with a high probability to convert. According to Google, in general, Performance Max campaigns can deliver up to 20% more conversions at the same cost.
Performance Max uses machine learning to continually optimise your campaign and has sophisticated levels of automation designed to make achieving your campaign goals easier. The machine learning models will optimise bids and ad placements across Google channels to drive whichever conversion value is more akin to your campaign goals. However in order for the models to work effectively, you must provide adequate data and inputs. One input is audience signals which include customer lists/segments and 1st party data. The more audience data you can provide, the harder the machine learning and algorithms can work to broaden your audience to consumers more likely to purchase and drive conversions towards your campaign goal.
Think, adding a larger amount of better quality fuel to your car is going to make it run better and faster for much longer.
However, if you don't have enough customer or 1st Party Data, you could expect to see a poorer performance from Performance Max and you may be better sticking to your usual search, display, shopping etc campaigns for the time being.
Benefits of Performance Max.
Reach more customers: Performance Max campaigns can help you reach more consumers across Google's various channels, including Search, Display, and Youtube. Performance Max campaigns use machine learning to optimise your bids, placements, and creatives in order to make sure that your ads are seen by the right people at the right time.
Improve efficiency: Because the machine learning supports optimisation of bids, placements, and targeting you could see an improvement in the efficiency of your campaigns and get more stretch from your budget. Google claim that PMax could help you avoid overpaying for clicks and impressions which may not be valuable to the brand.
Get insights: Performance Max campaigns can still provide you with information about the demographics of your target audience, the types of content that they are interested in, and the devices that they use so you are able to gain valuable insights about any potentially new audiences who are interested in your product. Google are slowing releasing more insights within PMax and allowing more granular reporting.
Powering Social Algorithms.
Much like Performance Max and Google in general, many social platforms are leaning towards algorithm driven campaign functions. Now manual targeting has taken a hit due to new privacy regulations, machine learning and algorithms can drive further reach and broaden your audiences based on data and realtime performance.
Meta has recently release a suite of new campaign formats labelled under the Advantage+ umbrella. Their aim is to support advertisers by leveraging machine learning to help reach more valuable audiences but with less set up time and more efficiency. In particular Advantage+ Shopping campaigns are uniquely designed to be one of the most effective and efficient social solutions for driving performance and conversions into online sales for ecommerce brands.
When comparing Advantage+ Shopping with Meta's original manual shopping ads, there are fewer inputs needed during the campaign build itself, therefore speeding up the process. Audience options are simplified and the management of assets should be more streamlines.
Meta give the example:
Instead of running many campaigns with varying targeting and creative setups, you can set up to eight Advantage+ shopping campaigns per country. This gives our system more opportunities to reach people likely to purchase your products.
However, in order to see audience-type breakdowns between new and existing customers and in order to drive more algorithmic efficiency, you need to be feeding custom audience sources into the set up of your Advantage+ campaigns. If you don't have a custom audience source, you are able to create one from your website, app, catalogue, customer list or any offline activity within the ad account set up within Ads Manager. Remember, the more information you can supply the platform the more learning its can take from the data in order to drive conversions. If 55% of your website traffic is now untrackable due to privacy changes, collecting 1st party or CRM data are really your only options.
Benefits of Meta Advantage+.
Streamline performance: You are able to combine prospecting and existing customer audiences under one campaign with personalised products from your catalogue and use Machine Learning to identify your highest value customers with minimal input.
Simplify campaign setup: Meta claims that by running Advantage + campaigns you can run fewer campaigns with the need to consistently adjust or refresh meaning less time needs to be invested in constantly setting up new campaigns.
Automatically test: Advantage+ allows advertisers to test up to 150 creative combinations and deliver the highest-performing ads based on the findings. You are then able to deliver your best performing ad variation to the highest value customers.
Machine Learning and AI can seem really overwhelming but it seems that advertising platforms are trying to make it as accessible as possible to make use of. We may not understand exactly how the algorithms work and we probably never will, but that shouldn't stop marketers from at least testing the capabilities these more automated campaign solutions. The solutions may not suit all brands, all year round, but they could certainly provide some incremental gain or provide useful testing insights when it comes to audience and creative.