It’s time to tear up the rulebook, says Marketo CEO Steve Lucas. He made the comments in an interview with Which-50 at the company’s recent annual user conference in San Francisco.
Almost two years after being acquired by private equity firm Vista, Marketo has invested hundreds of millions of dollars in product, engineering, and growth efforts to bolster its marketing engagement platform.
Central to that is a strategic partnership with Google Cloud, announced in August last year. Google Cloud, which is a current Marketo customer, will provide the infrastructure for Marketo’s applications in its public cloud as well as collaborating on machine learning products that will be built into the martech platform. The partnership also gives one of the world’s largest advertising companies an insight into the world of marketing technology.
In an interview with Which-50, Marketo CEO Steve Lucas said one of drivers behind the Google Cloud platform selection was “a deep conviction and belief that the worlds of ad tech and martech are converging.”
“We really wanted to be at the forefront of that and catalyse it,” he said.
“The Google decision was about how can we maximise the disruption in marketing, tear up the playbook, and deliver an entirely new set of capabilities.”
Lucas said it is a “tremendously bad idea to try and compete with Google and Facebook on advertising” as the two digital giants control over 90 per cent of digital ad spend in North America. Instead, Marketo’s products are aimed at “augment the effectiveness of ads placed on those platforms” based on the marketer’s own first party and third party data to place more specific ads.
Instead of competing with Google in the world of adtech, Lucas said it makes sense that Marketo would “partner with a company that is so massively influential in the world of marketing.”
“The reality is the relationship is fairly deep; we have a large commitment, not just around technology but engineering efforts and co-innovation and there is a well-defined innovation roadmap for years to come.”
Moving To The Public Cloud
Marketo is in the process of moving its applications on to Google Cloud.
Speaking with Which-50, Marketo’s CTO Manoj Goyal explained the decision to move from Marketo’s co-located data centres to the public cloud came about after the martech provider “almost completely re-wrote the stack” to handle an ever-increasing number of customer activities.
In the last 12 months, Marketo’s platform has successfully delivered 30 billion emails and captured 400 billion marketing engagements. That was not the case in 2013, Goyal said, and required a serious investment in a next-generation platform.
“We were trying to put this into our own data centres and what we were finding is that this infrastructure was getting very expensive in our own co-los (co-locations) and it was not fast enough,” Goyal said.
“We made a big decision that we are going to take everything, 10 years of investment in co-lo, and move it to the public cloud.”
They shopped around with Microsoft, Amazon and Google but ultimately chose Google Cloud based on scale and reliability, Goyal said. After announcing the deal last August, Marketo has begun migrating the first of its 5,000 customers and their data onto Google Cloud.
AI Product Development
Google and Marketo also have combined engineering efforts around machine learning and artificial intelligence.
Marketo has two AI products: ContentAI which developed in-house over a four year period, and AudienceAI that was built alongside Google and completed in a matter of months.
That speed of development is a key advantage of the partnership, Goyal said.
As well as using Google’s machine learning models and algorithms, Marketo has four of Google’s data scientists working at Marketo onsite with its engineering team.
The result is AudienceAI, announced at last week Marketo’s Marketing Nation Summit, which leverages Google ML to find lookalike audiences with marketers’ database to expand campaigns.
Marketo’s approach to AI is not to turn marketers into data scientists, but to build it into the user interface.
“I think that more and more AI is going to make mundane low-level grunt work decisions for the user,” Goyal said.
“People aren’t going to be geeking out on models, they will be prescribing goals and objectives and the system will take that into account and say ‘here’s what I recommend back to you.’ “