According to Adobe’s 2020 Digital Trends report, as many as 64% of large organisations use Artificial Intelligence to automate specific marketing activities. An area where the use of AI can significantly improve results is, among others, digital advertising. In this article, we discuss five key examples that show why you should consider this solution in your business.
Artificial intelligence in ad optimisation – how does it work?
Before we move on to specific examples, let’s try to explain why AI can actually help increase online ads effectiveness. Artificial intelligence can process and analyse vast amounts of structured and unstructured information, while standard techniques lack that possibility due to the size and diversity of big data.
As one of the machine learning techniques, AI allows us to build neural network models by processing massive datasets and capturing the patterns and correlations that are impossible to notice by humans. Machine learning also enables subsequent continuous refinement of the models and modifying their work to match new training data when it becomes available.
The above ML and AI features are irreplaceable when rules of operation aren’t clearly defined, and we need support in decision making or predicting occurrence of specific actions, having access only to large amounts of historical information.
Artificial intelligence in digital advertising — 5 key benefits
1. Enhance message personalisation
Nowadays, people are inundated with ads, and many don’t pay much attention to them. So, as an advertiser, what should you do to attract their attention? First of all, you should try to personalise your advertisement as much as possible.
How to do it? Use data about current and potential customers, their purchase history, browsing history or information from social profiles. Thanks to them, you are able to identify, among others:
- approximate age,
- marital status,
of your ad recipients. Then based on collected data, ads can be tailored to meet their individual character and preferences and thus be extremely personalised.
But that’s not all. Artificial intelligence is able to broadcast ads that will contain not only personalised messages but also images. ⬇️
Journal of Consumer Psychology conducted a study in which they tried to match images to personality types. Researchers, using artificial intelligence algorithms, identified 89 different features of images, including hue, saturation, colour variety, level of detail, the appearance of people in the image and more. The 745 participants in the study were asked to rate the images on a scale from 1 to 7. Then, they took a personality test that assessed them for: openness, conscientiousness, extraversion, agreeableness and neuroticism.
After the test, the researchers tried to relate picture preferences to personality types. Among other things, they found that:
- extroverts preferred simple images and images with people;
- open-minded people preferred images without people and in cold colours, such as blue and black;
- people with a high level of neuroticism liked calm and minimally stimulating scenes.
In the next study, the researchers noticed that the respondents preferred advertising images that matched their personality types. But importantly, the machine learning algorithm discovered that the relationship between personality type and image type impacted interest in a particular product. People not only preferred images that matched their personality, but they also showed greater purchase intent towards brands that presented images that met their preferences.
Let’s move to the next example of how message personalisation can help achieve business goals.
Some time ago, Starbucks wanted to become the most personalised brand in the world. Using machine learning, it created a real-time personalisation tool that integrated with customer account information, a mobile app, customer preferences based on other companies’ data and more. The result? Using AI-based predictive analytics, they were able to increase user interaction by 150%, which resulted in a 3x increase in net revenue per customer.
Another interesting example of how AI can personalise a message was recently shown by Hyperise, which created a personalised video with content tailored to the person watching it. Check it out to see how powerful the role of data and personalisation is.
The above examples are excellent proof that personalising your message allows you to build relationships with consumers at every stage of the sales funnel and increases their loyalty to your brand.
2. Targeting the appropriate audience
One of the biggest challenges of today’s marketers is targeting, which means addressing ads to the right people. After analysing a huge database, artificial intelligence can perfectly match the ad to people who are supposed to see it.
Based on previous campaigns, it can also segment audiences appropriately into those who are more likely to click on ads in a particular place (e.g. social media) or are more active during a specific time. Then, AI can tailor ads to them so that users are willing to perform a particular action. It not only allows you to increase the probability that they will click on your ad or purchase your company’s products, but it also expands your ad’s reach.
Targeting is used, among others, by a well-known social networking site – Facebook. Some time ago, the company launched a campaign that displayed ads for a cosmetic product that were targeted to the users’ personality types (based on their Facebook likes history). It turned out that, on average, about 1.5 in 10,000 people who viewed the ads bought the product. While 1.5 in 10,000 may not seem like an impressive result, it is worth emphasising that this personalisation increased the probability of purchase by 50%.
But that’s not the end of AI’s possibilities when it comes to targeting. How about using location data to target people near your store? Or maybe you want to see if sales of the assortment you have are affected by the current weather? Well, that’s not a problem.
Tools based on artificial intelligence will be able to find much more than only obvious correlations, such as between the weather and sales of sunglasses or umbrellas. Currently used neural network models can capture correlations that are not obvious to humans in any way – such as between a weather change and the purchase of certain food products, or between the weather and the time of day and the willingness to make an appointment to test drive a new car model. Additionally, all of this happens automatically, without human involvement.
3. Reacting in real-time
Nowadays, the world is rapidly changing, and your job as an advertiser is to react to those changes. In a constantly evolving market, you need to make sure that your ads are always relevant to your target audience. Artificial intelligence can tell you what is the interest in your product or service at any moment.
Let’s say you are the owner of an online gaming platform, and due to a pandemic another lockdown has just begun, it is undoubtedly an excellent time to increase the budget for your ads, as this is the time when consumer interest is most likely to increase. Using AI, you can make quick decisions to customise messages and change the direction of your campaign whenever needed.
41% of marketers say that AI and ML make their greatest contributions to accelerating revenue growth and improving performance
The real-time reaction also means adjusting ads to the relevant situation in which our potential consumers are currently at. For example, by working with GPS tracking software, such as Google Maps, gas stations can monitor distances taken by drivers since their last refuelling. Then, they send them reminders to use the gas station, additionally displaying ads with the latest promotions available in their stores.
4. Increase ROI
John Wanamaker, an American entrepreneur, said very relevant words: “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.”
And indeed, optimising the effectiveness of ads and then finding the elements that can improve results is one of the marketers’ most demanding tasks. Using artificial intelligence in this area can be extremely helpful because it can identify which ads and actions are most effective and what should be avoided, thanks to data analysis.
With the ability to process massive data sets, machine learning technologies are great for discovering revenue opportunities that budget and human assumptions may limit. Using AI in your advertising campaign planning will allow you to avoid wrong assumptions resulting in bad investments.
For example, if you sell video games, you may assume that they will mainly interest teenagers and their parents, completely ignoring older gamers. However, as it turns out, these types of assumptions can cost you a lot. Julie Shumaker cited this scenario at the VentureBeat Transform 2018 AI Conference. Advertisers may have particular goals, such as selling a $17 game license to a 22-year-old gamer. Their ads might not include a 65-year-old woman, but machine learning might reveal that it’s likely that this woman can spend about $4 over three days. And if it turns out that the cost of acquiring that consumer is significantly lower, we can improve the ROI (Return on Investment) of the ads in a non-obvious way.
Another critical issue that also contributes to improving ROI is the above-mentioned personalisation and targeting. By targeting the appropriate audience with a perfectly fitted message, your team can significantly increase the ad’s effectiveness and, therefore, improve ROI.
5. Plan further actions (predictive analytics)
It is worth mentioning that AI is also used to analyse demographics and activities on particular websites. The so-called predictive analytics, based on historical data, can predict data such as:
- bounce rate,
- page views,
- time spent on the website,
- Click-Through Rate (CTR).
It allows you to make better, data-based decisions and focus on the areas that get the best results.
To better illustrate how AI can help you customise your ads appropriately, let’s take a look at a study conducted by the British bank HSBC, in collaboration with Maritz Motivation Solutions. ⬇️
The company experimented with a loyalty program in which 75,000 customers were given points that they could exchange for various products and services. Some customers had their reward plan generated by artificial intelligence, while others used traditional promotional emails. The results indicated that 70% of the participants who finalised their purchase with a reward package preferred the recommendations generated by AI. This study shows that predictive analytics used by AI can be used to recommend specific products to customers before they realize they need such a product.
What about the actions that aren’t going too well so far? With the ability to identify past mistakes, we can analyse them and prevent similar errors in the future.
Thanks to these actions, not only more people can reach your content, but also your company’s budget will benefit from it as you will know which activities are worth investing in and on which you will likely just waste money.
Implementing AI solutions in your company
Do you need to increase the effectiveness of your digital advertising campaigns thanks to innovative AI models?
We have ready-made solutions that will allow you to improve your results immediately.