Do you really need digital identifiers like 3rd party cookies anyway or is the answer AI driven custom bidding algorithms?

A major problem for agency and client marketers is that the digital advertising landscape is complex and ever shifting.

Challenges facing modern marketers include the deprecation of third-party cookies, the quest for first-party data, the emergence of new channels such as connected TV and digital audio, or new metrics like attention, and the changing behaviour of target audiences.

I’d like to offer my opinion on at least 50% of that problem. That is, that the future of adtech on the buy side has artificial intelligence at its core. Advertisers and agencies are already using DSP custom bidding algorithms to sharpen and refine their targeting strategies, crunch through reams of data, and lift overall campaign performance.

This was evidenced in the recent Group M AI report, which noted brands that get it right should enjoy an advantage in consumer attention, consideration, and loyalty. The agency group also predicted that by 2032, AI will underpin everything from creative versioning to consumer purchase decisions, all in a privacy- first way.

The truth is that AI in programmatic is being embraced now because it enables agencies to drive better performance for their clients whilst being faster, and more efficient. In-housed clients of course, will also want all these benefits for themselves whilst making their digital media budgets go further.

The AI opportunity is here to fix problems in digital advertising and manage the level of decision-making, such as impression level custom bidding, that humans are not able to compute. In a world drowning in information and data, and with programmatic talent in Australia and New Zealand more in demand than ever before, AI needs to step up by eliminating tedious work, making automation easy and whilst delivering results 30-70% better than currently seen.

Done right, AI enables programmatic talent to focus on the other challenges such as measurement or verification, and more so opportunities which exist in the programmatic space at present.

Predictive AI is able to ingest log-level impression, attribution, and conversion data which is used to score impressions in order to customise bids and cherry pick only those impressions predicted to deliver the required budget with the best possible results given each advertiser’s unique KPIs and business outcomes.

All this means sophisticated AI can bid higher on impressions which are valuable to the KPI’s of a given insertion order before a consumer has visited your website or even heard of your brand or, on the flipside ensure you are not overpaying.

On that note, enabling AI driven, impression level custom bidding is something major DSP’s like Google’s DV360, The Trade Desk, Xandr, and MediaMath in Australia are fully supporting. They want campaigns to be successful and programmatic investment to continue to increase. Scibids, where I am proud to work, is an enabler of the full power of custom bidding at scale. One example of a bring your own algorithm (BYOA) solution available within a DSP is Custom Bidding, a very advanced product developed by Google with infinite potential. Despite Google’s announcement that they are delaying depreciating 3rd party cookies within Chrome browsers until 2024, there is still an immediate requirement for trading to be optimised against a brand’s true business outcomes.

Creating customisable AI-powered bidding scripts to bid exactly the corresponding $ value per impression to your insertion order will require a huge team of experienced data scientists to create and execute. Experts in this space know how to create scripts that speak the language of the buyer’s DSP of choice, of course to generate the best results at optimum cost.

At Scibids, in order to maximise clients’ outcomes, artificial intelligence is used to build and refresh custom bidding scripts at scale and often, rather than the market norm whereby data scientists push the models themselves.

Creating custom bidding scripts for programmatic media buying can see an immediate increase in the performance of both your brand and performance campaigns.

An example of this is a recent google case study featuring global beauty brand Charlotte Tilbury where they used Scibids to create custom bidding algorithms and deployable custom script in DV360 to optimise towards multiple KPI’s across display and video.

The challenge for the beauty brand, which other retailers selling high-value goods also face, is that it could not grow its media spend above a certain plateau without also growing the cost per customer acquisition beyond profitability thresholds.

This is because high price point products drive fewer conversions at a higher price, and this gives bidding algorithms less data to learn from. The brand needed a deeper data pool in which to perform. Our AI engine generated full-funnel, tailored algorithms multiple times per day that factored in relevant event data higher up in the user journey. This included valuing visits, clicks, and ad viewability as new input signals.

This approach enabled a sharp increase in conversions and lower cost of acquisition per customer for Charlotte Tilbury. Everybody is a winner - the brand, the consumer, and the agency.

That is the promise of AI, done right!

In the 2nd part of this 2 part series we will be covering off why custom algorithms are a crucial part of a programmatic traders toolbox when custom KPI’s such as Attention metrics need to be achieved, stay tuned!