Marketers must have a lot of different skills and wear many hats. While a broad spectrum of skills and activities is a big part of the allure of marketing today, it's also equally frustrating because there is so much to do and only so many hours in the day. Metadata.io has announced something that it sees as taking on the more technical, repetitive, mundane tasks, allowing marketers to get back to the strategy and creative parts of marketing. It's called the B2B operating system.
A much-needed technology?
At the launch, Gil Allouche, Metadata.io CEO, talked about the struggle to set up and run campaigns and all the manual testing he would do (he referred to it as "like a day trader on speed."
Part of this work is the repetitive, mundane tasks like setting up ad campaigns, running tests to see what's performing better, adapting as necessary, etc. While marketing automation was supposed to take over a lot of this type of work, Allouche argued that this didn't happen.
He brought in lots of technology to help, but like most marketers quickly realized that it's challenging to onboard all the different technologies needed, get them working together, and still have to go in and out of each one to do the work.
That’s why Allouche started metadata.io, pitched as a platform that automates paid campaigns. He wanted to build something to help marketers create "consistent growth that gets better over time and cheaper”, a platform that would operate, not replace, existing marketing and sales technology in a scalable, automated way.
But automating paid campaigns is only phase one of his wider vision to build an operating system for B2B marketers. Metadata's concept of a B2B operating system would automate a long list of "battle-tested" playbooks that can execute "millions of marketing tasks." It would handle all those mundane, repetitive tasks that marketers perform daily. But the vision is actually more than that.
So how would this work?
According to Logan Neveau, Senior Director of Product Growth for Metadata, this operating system would start with a list of playbooks the marketer can select from, such as pipeline generation, brand engagement, gifting, content marketing, organic social, email marketing, and more. All these playbooks are not built into Metadata itself (some are), but rather, Metadata integrates with other technology so that the marketer has one place to go to work.
Connecting your marketing technology with Metadata would pull the data from those systems into the platform. As a result, Metadata could cohesively look at all this data to provide the marketer with continuously updated insights and recommendations to improve campaigns and programs.
Neveau provided a few examples of how this works. For example, through data analysis, Metadata determined that 73% of a company’s upsells come from customers in two specific industries. The system determined that the company could improve upsells by 11% if it created upsell campaigns targeting these industries. The recommendation was to run an upsell campaign.
The marketer can choose to accept this recommendation and put the work required on "autopilot." Autopilot means the campaign will happen without any additional work required by the marketer - the system does all the work necessary to create and launch the campaign, including running multivariate testing to optimize the campaign.
Sometimes marketers will want to have a say in how a campaign will roll out but still have the bulk of the tasks handled by the system. They would do this with Metadata, said Neveau, through an UpVote flag. In this case, the system would take you through the process to set up the campaign. You would identify changes you want to make based on the system's recommendations (which could include everything from audiences, creative, channels, CTAs), and the system would then adjust and launch the campaign. As the campaign runs, it would continuously optimize the campaign to get the best results.
If this sounds good, but you're concerned about how the AI is making decisions, a decision tree shows you all the moves the AI made and why.
You could already automate and run paid campaigns in Metadata, with other types of programs to come - hence the integration with the other systems and the additional playbooks indicated above. Neveau provided the example of having Metadata create and publish blog posts, including the ability to pull text from existing content eg: website, call center transcripts, high-performing keywords, and headlines.
This is very much about data
Metadata has a solid client base today, which means the firm has a lot of historical data to train the AI and build predictive models. Allouche indicated that it has aggregated (anonymous) data from both first and third-party sources that the system could then use for its playbooks. That data grows even more through integrations with other systems.
Without all this historical data, this B2B operating system wouldn't be able to take away these technical, repetitive tasks because the marketer would still need to have a level of oversight and review.
It also means you have to know that the companies that Metadata works with today match your industry or company size, because you want the AI to compare your situation and campaigns with similar companies. There is, of course, the idea that it also pays to look outside your world and see what works for other industries.
Operating system or ecosystem?
Today it's common practice that successful technology software brands build ecosystems; the company provides an operating framework where other tech vendors could integrate to share functionality and data. Here you might think of the likes of Salesforce, HubSpot, Uberflip, and others.
So, is Metadata building an ecosystem? In a way, it is. It integrates other technology through Rest APIs and talks to those systems to perform tasks. If you want to execute programs that rely on one of these integrated systems, you are limited to what Metadata currently works with, or you take the heat for getting that integration complete. As an example, there are at least four gifting software companies I can think of - does Metadata integrate with all of them?
It's also pulling in data from those systems to give customers a complete picture of their customers - at least from the perspective of the tech that's integrated.
But Metadata is taking a more significant step thanks to the AI and the automation built into their system. The pitch is that not only can you integrate the technology and work from one screen, but you could get insights and recommendations to improve the performance of campaigns in all of those systems. And then, a good chunk of that work is automated for you.
I get this idea of an operating system, and it’s certainly something that many B2B marketers will appreciate because it gives more time back to strategy, research, and creativity that needs to happen to build the right experiences.
Metadata said the idea isn't to replace people but to take away the things marketers "didn’t sign up for.” I personally like the idea of an operating system that can handle much of the technical, mundane tasks that marketers do. I’d much rather design a campaign than do the work of implementing it.
But there are marketers, especially those interested in marketing operations, that did sign up for this work. In some ways, it seems like Metadata is built for companies with small marketing teams that don’t have the time to do everything, but it’s not - they recommend a spend of at least $20k a month across paid channels before you work with them.
This operating system is not going to support every marketing program, though. Content marketing has many moving pieces that require a unique set of features - many that aren’t automated or repetitive. Unfortunately, there isn’t a content marketing technology that supports everything you need for content marketing; I’d like to see one some day, but for now I’m not sold on the content marketing use case because it boils content development down to a set of tasks that can be automated, and I don't see blogging like that. Yes, I realize all kinds of AI tools today help you write blogs (and maybe Metadata is looking to acquire one), but I don't think an AI can build the quality content brands need. I'm also curious to see how some of the other playbooks "play out" in Metadata when they are done.
I look forward to seeing Metadata’s operating system for B2B marketers come to fruition, and I expect we’ll see others attempt to build similar models in the coming years. Maybe it’s the closest marketers will come to getting a single-window tool for all their hard work.