Adobe Advertising Cloud Search: Powered by Sensei
Adobe Advertising Cloud Search powered by Adobe Sensei provides a combination for marketers that is unparalleled in the industry.
Image source: Adobe Stock.
by Pete Kluge
posted on 03-07-2018
According to a recent report conducted by Forrester Consulting, 80% of retail marketing executives believe AI will revolutionize the marketing industry by 2020.
What is AI?
Artificial Intelligence (AI) and Machine Learning (ML) are top buzzwords in the advertising industry. But what are AI and ML – and what do they mean for search advertising? Let’s define terms:
- According to Websters, Artificial Intelligence (AI) is “the development of computer systems able to perform tasks that normally require human intelligence, such as decision-making…”
- Machine Learning is a “type of artificial intelligence (AI) that provides computers with the ability to learn without being explicitly programmed…computer programs that can change when exposed to new data.”
What does AI mean for search advertising?
Search advertising is complex. Search advertisers navigate Billions of data points and decisions. Thousands of keywords multiply out by campaign settings (i.e. match type, budget and inventory source) and dimensions (i.e. audience, device, geography and time) to create a staggering amount of data. It’s too much work for humans to sift through all of that data and make the best decisions on budget allocation to drive the best possible performance.
Additional layers of complexity abound for search advertisers:
- Sparse conversion data – Only a small portion of keywords searched by a small portion of users actually convert and generate revenue leading to a data sparsity challenge. How can an advertiser optimize a search campaign when there is limited conversion data to optimize with? How do you optimize for keywords or product groups for which there were no conversions or purchases?
- Constant evolution – The marketplace is continuously changing, which requires bid optimization throughout the day, the ability to predict and simulate performance in the future (tomorrow or next week) for planning, as well as long term forecasting for seasonality and promotions.
- Understanding relationships – There is a great deal of diversity and variance in performance across keywords. Understanding the relationships to help optimize bids as well as build the right account structures to deliver the best performance is a challenge.
AI helps solve these challenges by leveraging complex data sets to build predictive models and make decisions on bids and budget allocation to best meet advertisers’ objectives.
Adobe Sensei and Adobe Advertising Cloud Search
We are excited to announce that Adobe Advertising Cloud Search is now powered by Adobe Sensei. Adobe Sensei is the technology that powers intelligent features across all Adobe products to dramatically improve the delivery of digital experiences, using artificial intelligence and machine learning in a common framework. Existing Ad Cloud Search AI features have been approved by meeting the standards of the Adobe Sensei framework, and new AI features will consider Sensei framework during developed.
Adobe Sensei helps solve a myriad of problems for search advertisers by applying AI and ML to variables like bid optimization, forecast models, simulations and spend recommendations. All the while being transparent on forecast models and accuracy.
Adobe’s unique data access across the funnel solves data sparsity challenges
Adobe Advertising Cloud Search leverages data integrations with Adobe Experience Cloud, which Adobe Sensei can apply to bid optimizations and forecast models. The native integration with Adobe Analytics enables the sharing of over 2,000 site engagement and conversion metrics hourly. Engagement metrics like time on site and number of page views by visitor are powerful signals of consumer intent that Adobe Sensei can use to optimize for performance — even when no conversion or purchase data is available for a given keyword.
In addition, audience segments from Adobe Audience Manager, Adobe Campaign and Adobe Analytics are available to optimize search bids and ad creative. Adobe Sensei also considers dimensions like audience target, device, time and geography in its forecast models.
Adobe Sensei models operate across various levels of time from execution to planning
Adobe Sensei can optimize bids intraday and daily, allocate monthly budgets and generate simulations for planning of spend distribution.
Adobe Sensei runs models for clicks, cost and revenue for different budget scenarios. Adobe Sensei’s machine learning uses unique Adobe data to train the models and then allocates budgets across bid units to best meet advertiser budgets and goals, selecting the best approach and making tradeoff decisions as well as continuing to learn over time to improve the models.
An advertiser can run simulations to answer key performance questions like how to most efficiently maximize conversions or see impact of increasing budget on predicted outcomes.
Adobe Sensei analyzes and borrows from relationships to scale data for optimization
Relationships across keywords and bid units can tell us a lot about how keywords with sparse conversion data may perform in the future. Analyzing semantic clusters (keywords with similar meaning), performance clusters (keywords with similar engagement and conversion rates), account structures (keyword groupings) and product hierarchies (product group similarities) give AI powerful signals to optimize the models.
Adobe Sensei analyzes similarities and differences across keywords structures and hierarchies and then applies these learnings to forecast models.
Full model transparency with accuracy reports
Adobe Advertising Cloud Search offers total transparency into bid unit models and their accuracy. While AI may learn from complex data sets and build predictive models far more efficiently than humans can, ultimately human marketers are driving the strategy. With this in mind, the platform offers transparency into individual bid predictions and decisions as well as post-campaign accuracy reports to validate the models. Model transparency and accuracy reports are always available for all metrics, enabling advertisers to drill into individual days and metrics for detailed analysis.
On average, Adobe Advertising Cloud Search consistently sees 90-95% accuracy in its forecast models.
What does this mean for the industry?
If you’re picturing a robot sitting at your computer and taking over your job managing and executing search advertising campaigns, rest assured. AI simply reduces the amount of manual set up and optimization required of humans so they can focus more time on strategic decisions.
Ultimately, what we’re doing with AI is helping advertisers drive the best possible performance, learn from massive amounts of data and make complex decisions in a constantly-evolving marketplace. With Adobe’s unique access to data and audiences, Adobe Advertising Cloud Search powered by Adobe Sensei provides a combination for marketers that is unparalleled in the industry.
Topics: Digital Transformation
Products: Advertising Cloud