Advances in online ad management technologies mean it’s a full-time job to keep up with what is available, let alone understand when to use what technique to best effect! However, it’s worth investigating AI tools as recent advances in the technologies have made them better than ever, and useful even for campaigns generating smaller amounts of data. AI tools are privy to industry data that someone managing an account manually wouldn’t have. These insights can help the AI optimise campaigns effectively even when click and conversion data is low so it’s always worth running tests to see how they work for you. Testing techniques across multiple campaigns will also help the AI work effectively by providing more data so don’t be shy to test across your account.
Working with an expert team can grant you access to the secrets of best practice use of AI tools for marketing campaign optimisation so your campaigns deliver the greatest return possible. Recommendations from your agency on what to use when and why will depend on several factors. Your agency should consider the sector you work in and whether you’re trying to create leads or sales or grow awareness or engagement, how sophisticated your tracking abilities are, what your objectives are; are you looking to grow market share or establish efficient sales returns based on your capacity, and the targets you have set for marketing activities in every channel as well as the other marketing activities you are running.
Whatever your circumstances there will be opportunities to test for every brand.
Google Smart Bidding
Google Smart Bidding technology is Google most sophisticated, fully automated set of solutions that combine a set of signals to set precise bids for every search, display and video ad auction with the aim of improving campaign efficiency. Each technique available in the suite of solutions needs a certain amount of data generated every day to learn from but can increase campaign efficiency and deliver improved return on investment for most paid ad players.
The technology rapidly evaluates vast data sets to set optimal paid search bids for every auction. It allows paid search managers to work faster as routine tasks are automated and there are different technology options to choose from. You can set targets to control bid amounts, maximise the number of conversions within your budget, or maximise conversion value within a budget, or deliver the highest possible conversion value while maintaining a target return on ad spend (ROAS). Discover more detail about the Smart Bidding options and the circumstances that each one works best for in our blog: A Guide to Google ads Smart Bidding.
In summary you can set targets to:
• maximise conversions in paid search Google display campaigns or Video campaigns – useful if you have budget constraints and you don’t have a specific CPA (cost per action) or ROAS (return on advertising spend) goal or conversion volumes are low
• achieve a target CPA in paid search, Google display or video campaigns – good if you want to grow conversions while sticking to a CPA goal, or you want to decrease your cost per action goal value
• maximise conversion value within a set budget in search or Google shopping campaigns – good if you have budget constraints or don’t have a set CPA or ROAS goal and you’re focused on maximising conversion value (you need to be tracking conversion value as well as conversions) and don’t achieve many conversions overall
• target a specific return on ad spend (ROAS) goal in Google shopping ads, search or display campaigns – good if you want to deliver conversion volumes while sticking to a ROAS goal, or if you have scope to relax your ROAS goal to a higher value and grow sales. Perfect for bigger budget campaigns where you track conversions and conversion value with larger numbers of conversions each month.
As I’m sure you’re realizing using this technology effectively is complicated, so take a look at our blog on the do’s and don’ts of smart bidding to give you more of a steer.
We use Google Smart bidding for a leading bookseller and newsagent’s shopping campaigns which delivered well over 170% increase in sales volumes at the desired ROAS. This work put them in a good position to earn increased online sales during the coronavirus crisis while maintaining cost efficiency. You can read more about it in our case study.
Google Responsive Search Ads (RSA)
Google have developed machine learning technology that tests paid search ad formats to discover the best performing combinations of headlines and ad descriptions. The technology works out which is the right combination of headlines and text for a particular customer and can appear against more search queries. They also provide feedback on the quality of your ad copy and headlines to help you improve your copywriting.
Google Dynamic Search Ads (DSA)
Dynamic Search Ads are a great way to fill any gaps that may be present in your keyword set, which is inevitable given that users are always changing the way they search. This is done by Google crawling your website and identifying searches which are relevant to the content on site, the feature automatically determines the landing page and generates ad headlines based on the users search and the content on the landing page. Descriptions still need to be written by the account manager, and there is complete control over which areas of your website you target with this feature.
Google Dynamic Display Ads (DDA)
Dynamic Display Ads are populated with content, text and imagery all relevant and personalized to an individual user. Signals of intent are often used to trigger what to display within a banner, with the content being pulled from a feed such as a product feed. Often in retail it’s the product a customer has viewed on your website, or in travel it could be the holiday destination they have been researching online, that gets populated in the banner. They are perfect for finding people searching for what you have and are good for retailers with large product ranges that are constantly changing which makes dynamic display ads very powerful.
Google Responsive Display Ads (RDA)
These ads work in a similar way to RSA but in display campaigns on the google display network and can boost conversions. They work especially well if you have HTML 5 ad assets and you can test multiple copy and visual combinations so that through the machine learning google serves the right ad to the right person.
DBM or Google Display and Video 360
DoubleClick Display and Video 360 is the consolidation of DoubleClick Bid Manager, Campaign Manager and Studio created in July 2018. The suite of tools allows advertisers to plan and buy display and video campaigns, design and manage creative and overlay audience data to enhance targeting and optimise campaigns all in one place. This programmatic advertising platform interacts with over 100 ad exchanges or networks (1 of which is the Google Display Network) to cost effectively buy available inventory from over 1 billion sites and allow experts to run and optimise campaigns cost efficiently.
This technology gives advertisers scope to extend their reach beyond just the Google Network and track conversions accurately, so sales or leads are not multi counted if several networks are involved in a purchase journey. You can also see what was involved in a sale journey to better understand how your campaign is working even if an ad is not clicked. This attribution technology is important for understanding the role of each type of ad (search ads included) in a sale to achieve maximum conversions and goes beyond last click sale attribution.
Your own customer data can be overlaid to enhance targeting, and there are multiple creative formats available to showcase your brand or product.
As in Google search AI technology the technology offers automated bidding technology to reduce manual interventions and tasks and deliver increased cost efficiency depending on the goals you have decide upon. The machine learning adjusts bids based on a wide range of signals some unique to the user viewing the page, some dictated by your campaign set up and overlays.
The goals available are:
• Minimise cost per click (CPC) or cost per action (CPA) while spending the budget. It needs at least 100 clicks and 10 conversion a day to learn effectively so is good for larger advertisers.
• Meet or beat a CPC or CPA goal of x. This technique focuses on the goal set rather than spending all the budget available. It too needs 100 clicks and 10 conversion a day to learn effectively.
• Maximise quality video impressions. This technique is great for awareness campaigns ensuring more of your total video is viewed
• Optimise for viewable cost per thousand impressions. This too optimises for viewed ad impressions.
Deciding which strategy is right for your campaigns will depend on several factors as with search ads. Size of budget, rate of data creation (clicks and conversions), campaign objectives and creative format will all influence your agency’s decisions. Testing the AI technology performance against existing campaigns is always recommended but it does take time! Campaigns need to be running for at least 4 weeks to allow the technology time to gather data and learn from it.
• Like any machine learning these techniques need a certain amount of data to use to learn and improve. It’s important to realise that campaigns need to be live and learning for a decent amount of time to reach optimum performance. This might be less than a week for some big shopping campaigns or a month or more for small campaigns generating less data. Hold your nerve!
• Set realistic target Cost per action or Return on ad spend metrics. Google can recommend suitable targets or previous activity might give you a steer on this. The technology is clever and will improve campaign performance in many cases but it’s not magic – once the strategy is able to match previous performance then it is time to make small changes to targets to make incremental improvements.
•Long delays between first contact and conversion can delay the rate at which your campaign learns – bear your typical sales cycle in mind.
•These campaigns need expert set up. It’s complicated stuff and making too many changes too frequently can mess up the learning and spoil your potential results – again, hold your nerve and listen to your experts. Your campaign may just not have been live long enough in a new set up to optimise performance yet.
• Make sure you run tests of the new technology in a valid way – run experiments comparing old techniques with new on 3-5 campaigns minimum; run tests for at least 4 weeks, maybe a lot longer if your sales cycle is long; split your budget 50:50 between the old techniques and new and make sure you run tests in campaigns with enough conversions to split the campaigns in two. And remember to compare the same metrics in each test to evaluate the winner and factor in conversion lag.
If you’d like to know more about how AI could supercharge your campaigns please get in touch.