Does Adgorithms have the secret sauce for artificial intelligence marketing?
- It's a common assumption that AI isn't ready to drive marketing. But are we further along that we think? Based on Barb Mosher Zinck's review of "Albert," the artificial intelligence tool from Adgorithms, we might be turning a corner.
Now before we talk about Albert, let’s have a giggle because we already have Einstein and Watson, do we really need Albert? If you looked at what he does, you might think so.
Albert is the brainchild of Or Shani, CEO and Founder of Adgorithms. Shani worked in marketing for large brands doing things like business intelligence, media buying, analytics and data analysis. He commented that online marketing looks so sophisticated on the surface - but it’s a lot of hard manual work. And he said it’s hard to do a decent job. That’s when he thought about starting Adgorithms and simplifying and automating the entire process.
Shani sees two primary roles in marketing. There’s the creative part where you need to develop the right images, slogan, color, font, content, and so on and it’s very much an ideation process. It’s something you can’t automate. But then there’s the execution and big data components. That work can be automated in Shani’s opinion. He acknowledged that some things are a bit challenging to automate, but they figure it out.
The things that Albert can do
Initially, Albert was built to support display advertising. Over time, they added Google, Facebook, and email into the mix. So, what exactly does Albert do?
Albert does your media mix for you, he runs cross-channel campaigns, tests and optimizes and analyzes doing more than giving you insights, but adapting automatically based on what he’s learning.
Shani likened Albert to the Google self-driving car. It may not be beautiful with all the bells and whistles of a Mercedes (think Adobe or Oracle), but if you want the features it offers and want that automation, then you want to consider Albert.
You create your campaigns, assign a budget and a period and Albert figures out the best mix of advertising and communications across all your channels. He creates the message, sends the message, tracks the message. He learns in real time what is working and what isn’t and he adapts the mix accordingly.
There are three parts to Albert’s user interface (which is cloud-based): the dashboard, the campaigns and the creative. In this view, the marketer builds creative, hooks up the creative to a campaign and then sees the results in the dashboard.
The dashboard is your real-time view into what Albert is doing/has done. It shows you the breakdown of activities he performed including cost, impressions, clicks, conversions, revenue, ROI and so on. You can look at the breakdown of conversions such as calls, purchases, leads. There’s a channel breakdown view, and you can filter, mix and match the things you want to see to get a better understanding of what’s going on.
The dashboard in Albert is built not just to show you how specific campaigns are doing, but also to show you the effect of those campaigns on other things like social and your website.
According to Shani, Albert doesn’t hide anything. He brings everything to the surface, so you can see what he’s doing and the results of those efforts. Albert will also give you “Insights.” Albert runs a bunch of reports for you, processes the information and then shows you the things that are the most interesting and relevant to you. He also recommends what to do with these insights, including adjusting creative, channels or your target audiences. So, not only can you look at raw data in the dashboard yourself, but Albert also looks at it all and makes recommendations based on what he’s seeing.
Setting up a campaign is straightforward. You configure your audience segment targeting by existing contacts (from your CRM), visitors (from a cookie), look-alikes or new visitors. You can narrow this audience down by a range of attributes. Other than identifying the creative, budget, time period, Albert does the rest of the work, creating the campaign behind the scenes. He tries things in small increments - mini campaigns - learns and makes changes as he figures out what’s working.
When you add your creative to Albert, you can use all the normal assets: text, images, video, headlines, email, and other things. Messages are broken down into concepts, and you can specify which concepts run in which steps of the customer journey (yes, you configure Albert to understand your customer journey paths and Albert can run campaigns using that path, presenting unique concepts at every stage, even unique for each visitor.
The whole picture in one place
Shani and CTO, Tomer Naveh, gave me a demo of Albert, but the techie in me wants to know what’s going on behind the scenes. On the backend, it connects to all your accounts: your CRM, social media, Google Adwords, Google search, ad exchanges, email, and all the rest.
Naveh explained it by saying that some people say campaign execution involves a lot of intuition - when to be more aggressive, when to be less aggressive, what websites to target, what keywords to target. But when you sit down with them, and you interview them, observing what they do you find that this intuition can be broken down into pieces that the machine can imitate and repeatedly do, and then get feedback and correct itself over time.
He said people might think there’s one magical algorithm that you plug in, but there are tons of micro decisions and types of decisions that you can break down into small, well-defined problems and figure out the algorithm for each problem. Then you have algorithms to tie everything together. Once you teach a machine to make those decisions and it does it with high frequency it becomes so much more efficient.
Needless to say, many of those algorithms are proprietary to Albert. And that’s about as much as I can tell you about how Albert works on the backend.
So, I was impressed. So much that I almost wondered why some big company (like Adobe, or Oracle, or IBM) hasn’t snapped Adgorithms and Albert up. There’s some pretty heavy lifting happening behind the scenes to make Albert as smart as he is. Maybe Salesforce would be interested.
But it’s not a secret sauce in that you can see all the raw data and you can crunch all the numbers yourself. And maybe you would do the same things in the same combinations as Albert, but certainly not with the frequency or speed. And there’s no way you could do it to a depth that Albert does, narrowing down customer segments to as a close to a 1:1 experience as you can realistically get.
Adgorithms has done something cool here. But what I’ve been finding since I started covering the topics of machine learning and artificial intelligence in marketing is that there are more martech vendors out there figuring this out. Look at Demandbase’s recent announcement of AI-based personalization. Or Oracle’s announcement of AI tools for marketers. Or how about IBM’s new AI-powered Marketing Insights. Things are happening faster than we think. As long as the transparency is there to understand how the AI works, then this is a good thing.