“The Sell Sider” is a column written by the sell side of the digital media community.
Before she was CEO of The Weather Company and GM of IBM Watson Advertising, Sheri Bachstein was a field producer for a Weather Channel storm-tracking team. During her time chasing storms, she learned how reporting can save lives and make sense of the destruction caused by extreme weather events.
That experience still informs Bachstein’s thinking as she leads her company into a post-cookie world.
As a large publisher, The Weather Company has access to a large amount of first-party data, and IBM Watson Advertising has been making big investments in AI technology. Bachstein’s goal is to use both to monetize the creation of quality content rather than pursuing short-term gains in ad revenue.
Bachstein talked to AdExchanger about The Weather Company’s decision to show fewer in-app ads, lessons learned from introducing a subscription tier and how IBM Watson and The Weather Company are using AI to mitigate bias in advertising.
AdExchanger: The Weather Company is cutting in-app ads by 42%. How will this change the experience for users and advertisers?
SHERI BACHSTEIN: Over the past few years, the industry got out of balance with monetization versus content. I challenged my team to get us back in balance. The team spent four months testing variations and looking at consumer satisfaction versus monetization, which resulted in a 42% reduction in our ad footprint.
Users will see fewer ads on our platform, so they get the information they want without clutter, while advertisers will have more space and richer formats for their messaging – and less competition for the KPIs they’re driving.
What lessons have you learned from introducing a subscription tier and how do you plan to develop that aspect of the business further?
There was a big movement toward subscriptions with COVID, and there are statistics showing that subscription revenue is going to rival ad revenue.
We’re about two years into our subscription model with around a million subscribers. But things can stall after so many early adopters. So how do you get over that? Focus on acquisition and retention, although retention is probably even more important than acquisition.
We’ve built our subscription business side by side with our users. We did a lot of testing and ran focus groups. We asked users what they would be interested in, what they would pay for, what pricing. That’s been very beneficial.
Pricing is really important. Make sure you launch with the right price, because it’s hard to change your price.
How are you combining your first-party data with your AI tech to improve your subscription model?
We’re applying AI to determine which users have more propensity to subscribe and what messaging works for them. We’ve used our accelerator tool to run targeted promos on our platform.
We’re also using AI tools around propensity to churn – who hasn’t been back to the app in a while, even though they’ve paid for it, and how can we give them something of value to get them to renew?
Subscription models can limit access to information. You know weather reporting can save lives from your days as a storm chaser. How do you make sure you’re not putting content behind a paywall that people need to see?
Our job is to give people information to keep them safe. We would never put any of that information behind the paywall.
We put advanced weather data, like advanced radar information, behind the paywall. We have a lot of users who are interested in things like that, me included. Rather than gating content, we decided to create more value for the money. We look at opportunities for added services that a niche group of people are interested in and willing to pay for.
How else has AI helped improve The Weather Company’s ad business?
We’ve been part of IBM for six years, and it’s given us this amazing technology playground. We always used AI in our forecasting at The Weather Company, but this allows us to apply AI to our advertising.
Weather can influence behavior or buying patterns. We’ve created triggers and segmentation and a couple hundred different signals related to weather. Now we’re working with companies to distribute this into their platforms, including our DSP, SSP and publisher partners.
Signals like weather are going to be critical as privacy laws continue.
How can AI help the ad industry solve the identity and measurement problems it’s facing as cookies go away?
Cookies tell you what happened in the past – where a consumer has been. With AI, we can see what’s happening in real time and what’s going to be happening in the future. We are becoming more predictive around consumer behavior. Our solutions are doing that without cookies or traditional identifiers, using AI to determine signals from massive amounts of data.
How is IBM Watson using AI tools to address bias in advertising?
Bias is a systemic problem, and our research has found it exists in advertising. I don’t think anybody’s surprised about that. But how can we mitigate it to the benefit of the consumer and also the marketer? Can we achieve their KPIs with a mitigated advertising solution? We’ve found that [bias mitigation] opens up great opportunities that advertisers were unconsciously eliminating.
We’re looking for partners that will share their data so we can continue our research and analysis. We’re also creating an open-source toolkit and a library of tools that anyone in the industry can use.
This interview has been edited and condensed.
