Another day, another generative AI (ad)venture.
Google unveiled a marketing product Wednesday that aims to shorten the time marketers spend launching personalized campaigns. Making it easier to create many audience segments and creative versions could reduce weeks of work to hours.
To build the generative AI product, Google Cloud’s BigQuery collaborated with enterprise generative AI company Typeface to handle the creative assets and mar tech company GrowthLoop (formerly Flywheel) to handle audience building.
When advertisers build a marketing campaign in Google Workspace, generative AI will assist them in creating customer audiences and content and measuring and optimizing performance.
To take this work off the marketer’s plate, the gen AI tools need a deep lake of the brand’s first-party customer data to draw from, which is where Google Cloud and BigQuery, Google’s cloud data warehouse, come in. Based on BigQuery data, GrowthLoop creates and refines audience segments. All the information GrowthLoop gathers about audience segments then feeds into Typeface.
Typeface generates personalized content for each of the audience segments that matches the brand’s voice, style and tone – and makes varied versions for blogs, landing pages, social media and other channels. A virtuous “closed-loop marketing cycle” ensues, as marketers can use analytics data from campaigns to “fine-tune” both the segments and content, said Vishal Sood, head of product at Typeface.
Say a marketer wants to target every customer who hasn’t purchased in the last 180 days or everyone who’s likely to buy a pair of jeans in the next week. When they ask GrowthLoop to find a given audience, GrowthLoop generates the SQL, runs it in BigQuery and presents the marketer with an audience report showing the makeup of the audience. Marketers can generate multiple segments that address customers differently, depending on where they are in the customer life cycle, according to Chris Sell, GrowthLoop’s co-founder and co-CEO.
Marve, GrowthLoop’s generative AI feature, can iteratively produce segments and customer journeys. If a marketer asks for, say, email and push notifications for 15 audience segments in five cities likely to buy electronics, the gen AI delivers.
GrowthLoop also creates what it calls “audience persona prompts,” Sell said, which draw on customer ZIP codes, purchase category preferences and whether they interact more with the mobile app or email. GrowthLoop supplies these prompts to Typeface to custom-tailor creative assets. For instance, Typeface might generate 20 creatives based on the top cities where the users are located.
If customers wish, GrowthLoop can hold back randomized samples from each audience to measure which audiences produced statistically significant revenue lift, more mobile app log-ins or whatever metric customers care about. “Being able to identify the diamonds in the rough becomes an automated process within GrowthLoop,” he said.
But despite all the efficiencies the product introduces, marketers need not fear for their jobs. Most marketer roles “have a very, very long backlog today, and this [tool] just helps you get ahead of that,” Sood said.
Sell, who started his career as an email marketer at Google, left the field to found a tech company in part because of the frustrations of playing project manager and doing the endless “Excel spreadsheet dance.” He thinks marketing jobs will shift to become more interesting, with marketers gaining a greater ability to execute on their ideas. “As a marketer, if you’d told me this future was possible, I would have said, ‘I might want to stay in marketing,’” he said.
Wednesday’s collaboration and tech integration announcement is only the latest example of Google’s investment in generative AI. Although Google has used AI tech for a long time for tasks like translation and speech recognition, it’s been perceived as being slow to board the generative AI train compared to Big Tech and AI startups. For instance, its chatbot, Bard, came out months after ChatGPT. Google is now striving to catch up to rivals like Microsoft, OpenAI and Meta with a blaze of AI announcements.
In addition to this collaboration with Typeface and GrowthLoop, this year, Google’s rolled out myriad AI tools, such as AI-powered YouTube video campaigns; Imagen, a text-to-image diffusion model; Vertex AI, a cloud-based generative AI tools suite; PaLM LLM, which is integrated into Google products such as Maps, Docs, Gmail, Sheets and Bard; and Duet AI, a “copilot” generative AI tool for work tasks like summarizing meetings and sending follow-up emails.
Now, Google is betting marketers will want to get in more on the AI fun. As the song goes, “A little less conversation, a little more action, please.”