Testing, Testing: This Startup Uses AI to Make Websites More Engaging

WEVO helps marketers analyze their messaging.

Written by Nona Tepper
Published on Sep. 27, 2019
Testing, Testing: This Startup Uses AI to Make Websites More Engaging
WEVO
image via shutterstock

Most digital marketers have little insight about what people will think of their campaigns before launch. Will they love the photos? Does the headline make sense? Should you put pricing information front and center?  

Nitzan Shaer founded WEVO in 2017 to help marketers analyze their messaging before going live. The Boston firm uses a mix of human responses and artificial intelligence to review why a specific page on website isn’t inspiring visitors to buy products, request information, or sign up for company newsletters. The platform then provides suggestions for how to improve it, with recommendations ranging from layout to copy to design elements.

WEVO charges a subscription fee to review websites. According to Shaer, customers’ conversion rates rise between 50 and 200 percent, on average, after the firm reviews their digital marketing campaigns.

 

What is a campaign, exactly?  

A campaign is a term describing a coordinated marketing effort to achieve one specific goal — often positioning a product or service in the eyes of potential customers. Campaigns can take many forms, from targeted emails to ads on Facebook or Google, but WEVO currently focuses primarily on web pages.

For example, if the campaign is a web page designed to sign people up for tax preparation services, all the elements — including the text, the photos and all the visual design elements — should come together to promote a single goal: getting people in the door during tax season. 

Eventually, WEVO hopes to move beyond single webpages, and analyze the effectiveness of marketers’ videos, emails and more before sending them to customers. 

We see ourselves as part of a first step in a movement toward pre-live optimization.”

“We see ourselves as part of a first step in a movement toward pre-live optimization,” Shaer said. 

Shaer came up with the idea to develop WEVO when he was working as head of mobile at Skype, which was located in 173 countries at the time. Users ranged from grandparents who wanted to videoconference their families to businesspeople who wanted to seal deals across the world. 

Curious how Skype marketers spoke to such a diverse audience, Shaer asked his friends how they determined their messaging. He realized they were frustrated: Individuals spent so much time designing and iterating on campaigns that flopped and, since they had no idea how a message was going to perform until after it was rolled out, there wasn’t much they could do to avoid the expensive and time-consuming process.

 

What is human-augmented AI?

Inspired, Shaer developed WEVO, which he said uses “human-augmented AI” to review marketing campaigns before they go live. 

Customers like Bentley University in Massachusetts start by submitting a sample webpage to WEVO  — marketing, say, a primer on its undergraduate business degree program. Bentley outlines for WEVO what type of students they’d like to enroll in these classes, choosing from the 170 filters that can segment users by such factors as intent to apply, or whether they are considering other universities.

Once Bentley identifies its customer base, WEVO determines the statistically relevant number of reviewers needed to look over the university’s sample campaign. WEVO partners with companies in the crowdsourcing space — Shaer declined to specify which ones, but said their partners give them access to about 30 million people. 

We’re the first company to ... predict how well [a] website will be converting.”

Normally between 150 and 300 people review a webpage, Shaer said, offering comments along the lines of, “I see that the university can help with cost, but it doesn’t mention a price” or “I like the quotation in the front and the man graduating picture because it shouts success.” 

Reviewers also talk about what they expected to see on a webpage, versus what they actually saw. 

Their responses are then compiled and fed into WEVO’s machine-learning system, which pairs human feedback with historic website testing data to build a predictive model. Shaer said the firm has acquired A/B test pages that it uses to compare sites through various partnerships.

“To the best of our knowledge, we’re the first company to look at a set of information about a website and ... predict how well that website will be converting, as well as what are the hurdles that are preventing people from converting,” Shaer said. 

 

Bentley university
image via bentley university

Clarity, credibility and relevance are key 

When studying why students were not applying to Bentley’s undergraduate business degree, WEVO analyzes its marketing webpage for clarity, its appeal to relevant students and credibility. 

“Do people understand what’s written on the page?” Shaer said. “Or is it marketing gobbledygook that people inside the company understand really well, but the external consumers don’t?” 

Sometimes, Shaer said, it might just be the way the messaging is presented. In that instance, WEVO could recommend reformatting the text on the page, adding bullet points, or changing the tone. 

The firm also analyzes how relevant the marketing campaign is to the target audience, and if the webpage’s user experience drives them to apply to the school. 

In addition to providing a written report about why customers aren’t converting, WEVO creates a “geographic heatmap” image that uses green to highlight the good, and red to highlight the bad, parts of a website.

Shaer said the heatmap’s technology is “influenced by the predictive model” and differs from how firms traditionally created heatmaps, which was by tracking mouse clicks and movement on a page. According to Shaer, WEVO’s heatmap offers far more actionable insights than the traditional variety since it’s based on feedback from people who fit the marketing campaign’s mold.

“We could point, for example, to a certain image that people are offended by or that doesn’t resonate with them,” Shaer said. “We could pinpoint a quote from a customer that just doesn’t jive well with them. We could highlight a message that is resonating very well with the target audience and saying, ‘This is why I’m going to sign up for the product.’”

 

heat map screenshot
image via wevo

Next steps: Expansion and new features

WEVO’s reviewing process can take anywhere from a couple of days to a couple of weeks, depending on how hard it is to obtain the target test audience, Shaer said. For Bentley, it took less than 20 days to review the school’s webpages, and WEVO’s insights helped drive a 151 percent increase in the number of applicants for the school’s business program. 

Stories like Bentley’s have caught the eye of investors. 

In late July, WEVO received $2 million in venture funding, bringing total investment in the firm to $3.6 million. Shaer said the firm plans to use it to add two salespeople and five developers to its 14-person team over the next six months. 

“We’re adding multiple members to our development team to strengthen and deepen our tech and to expand into new areas,” Shaer said, adding that WEVO would make new feature announcements in the coming months. 

The firm plans to make “deeper inroads” in the consumer product good and financial services sectors. 

Long-term, Shaer said he’d like to make WEVO available for consumers. As an example, he said that, before someone starts a Facebook fundraising campaign for a nonprofit, WEVO could give them feedback about how to structure their ask to maximize the number of friends reaching for their wallets. Shaer said he believes consumer optimization technology will be a standard in five to 10 years. 

“Sometimes, the people that are the creators, innovators, that have these awesome ideas may not be the absolute convincer,” Shaer said. “We can help them provide tools to spread ideas in a better way.”

Hiring Now
CarGurus
Consumer Web • eCommerce • Software