Original article was published by Jason Pelish on Artificial Intelligence on Medium
Two New Metrics for the Age of AI Marketing: AI Consideration and Recommendation
Robots are doing stuff behind our backs, stuff we need to keep track of. The problem is real, here today; it’s not some imagined science fiction horror. These robots are hiding, hiding what they do, and who they do it to. They’re making choices and guiding decisions right now, and there’s no window for the public to see inside their minds to know why. Marketers need new metrics, new ways to see and measure what artificial intelligence is doing. These robots aren’t doing rocket science, so breaking down what these new metrics need to represent is pretty straightforward.
In the last few years, technology got way ahead of our ability to keep track of it. Online metrics, ways to measure a brand’s performance online, have some growing up to do. New devices that let people interface directly with artificial intelligence aren’t sharing what they’re doing with online marketers — yet. Online marketers are starting to ask questions about what’s happening inside these devices because measurements are valuable to know. Marketers need to measure how much AI is considering and recommending brands’ products and services. New metrics are needed that reveal behaviors in the cloud, high above the classic online conversion funnel.
What are the classic online metrics and why are they lacking today?
Online marketers have had 3 basic metrics to gauge performance of campaigns: Impressions, Responses and Conversions. Here’s how they work: When someone sees an ad it’s counted as an Impression. When someone clicks on an ad and becomes a visitor to a site it’s counted as a Response. When that visitor performs some action like buying a product, or subscribing to a newsletter that’s counted as a Conversion. Together these metrics make up a “conversion funnel” which refers to the expectation of numbers tapering down with each step.
These metrics are lacking today because they do not account for activities external to the site that aren’t Impressions. If a tree falls in the forest does it still make a sound? If a human being never is shown or becomes aware of a brand, its products or services is it still an Impression? No. Not really. With a classic Impression there’s still a chance of the brand making an impression on the audience. They still see an ad or a mention. Just like the expectation with a television or radio commercial, the person seeing or hearing it becomes a “brand aware.”
What are the new metrics of electronic agent consideration?
There’s something else that emerged yesterday that’s going to matter even more than Impressions in the future: how many times artificial intelligence considers your brand’s products or services when it compares competing products or services in the course of fulfilling some request for its user.
Following that new metric of artificial intelligence consideration is another important metric: how often your brands products or services are recommended by artificial intelligence.
TABLE OF METRICS
AI Consideration — AiCo
AI Recommendation — AiRe
Why don’t we see these metrics today?
Most online marketers are familiar with web traffic measuring products like Google Analytics. These products are designed to show human traffic and exclude bot traffic. If you examine the number of hits in a site’s web logs versus the numbers shown in Google Analytics you quickly learn how good of a job Google Analytics does separating the bot traffic and hiding it.
When bots visit your site that isn’t bad though. Bots are good because it means they’re paying attention to your content. For Google to include links to your site in Google Search it has to send its web crawler bots to your site. For Facebook to know how to show a link correctly it has to send its bots to a site to get the images and title of the link. If there’s a lot of social sharing from the site then the bot traffic will be higher than at a site that doesn’t.
So now, if a bot crawled your site in real-time, to fulfill a user’s request, you might not see it in conventional web analytics systems intended to be useful to marketers. It might not get the link from Google Search or another external search engine, but instead have the link cached, or the entire page content cached, on its own database. That might make it able to fulfill requests involving information your brand is publishing on your website without there ever being a search that Google knows about or a request for page content that could ever be seen in a weblog. These activities might be taking place multiple times a day from a page crawl done weeks ago.
As an online marketer who has developed proprietary analytics systems I know how difficult it is to keep up with the bots. Some of the people who design them often make them look and behave like human users because they’ve run up against web systems that discriminate against bots.
I designed tools, like Google Analytics, intended to filter out the bot traffic, treating it as garbage essentially. It’s always been nice to see the bots show up when something new gets published or shared, but they were never included in our performance counts for campaigns. They were noise.
Now bots matter more. Content marketers are beginning to practice structured data marketing, making sure their content is marked up for the bots to understand. Using structured data, included invisibly inside a web page’s code, makes the difference between being included in a search today or not. Tomorrow it’s role will grow as artificial intelligence’s role considering brands and recommending them grows.
Why will AiCo be the most important metric in the future?
Marketers want to know how often bots get asked for things the brand sells. Marketers need to understand how often bots are CONSIDERING the brand, the brand’s competition, or something that isn’t relevant competition at all. Marketers need to know how often bots are RECOMMENDING the brand from among a list considered. Marketers need to know so they can improve these rates.
How will businesses compete for greater AiCo?
Businesses will compete for greater performance with these metrics by communicating better with the bots. Businesses who can understand the limits of artificial intelligence in their market, against their unique business rules will thrive. It’s as basic as asking “what are these bots trying to do for their users and what can I do to make sure they see me as the best choice”?
Search Engine Optimization isn’t rocket science. It just involves disciplined work. If you put yourself in the role of the Search Engineer, trying to deliver better, more useful search results to the search engine user then it’s clear. They need to establish credibility. So do the bots. They both use the same information, information that brands give to them. Some marketers, once they understand how structured data works, think it’s easier than SEO, more straight forward.
How are these new metrics going to become available to marketers?
There’s no standardization between manufacturers of interactive AI countertop appliances and smartphones. I can’t see, right now, how many times Siri compares my services to my competitors. I can’t see if Alexa recommends someone get information from my site, like a phone number or working hours, pulled from a cache loaded weeks ago. These manufacturers don’t publish the data around these activities in any standardized way.
I hope this article highlights a need for this vacuum to be filled. It would be great if Voice-First application developers exposed these metrics, AI Consideration and AI Recommendation, to their business audience.
— — — — — — — — — — — — — — — — — — — — — — — — — — — –
Jason Pelish is an online marketer, a former analytical chemist, LIMS programmer, and a creator of WordPress plugins.