As we wind down 2023, there’s talk everywhere of generative AI and how it will fundamentally alter the world as we know it; but how does that translate for your corner of the business world? Is TikTok something you need to take seriously? (Is it time to dance?) We tapped into the research of seven HBS faculty members and a bonus Blavatnik Fellow to bring you the latest ideas on these topics and some time-honored challenges such as delighting (and keeping) your customer. Want to dig into the nitty-gritty of retaining workers in a post-pandemic world? We’ve got that covered too, along with more news you can use to bring yourself and your business up to speed in the year ahead.

It isn’t news that the fundraising climate has cooled considerably in the past year. But as a startup advisor, angel investor, and former executive at several tech companies, Senior Lecturer Julia Austin has a more nuanced understanding of this reality than most—and a few pointers for anxious founders to keep in mind.

“It’s still unpredictable, hard, and illogical in some ways,” Austin says of the current environment. “What’s different is that valuations have gone down. We’ve seen some odd behavior in venture capitalists trying to revalue a company they’ve already backed. There also are companies raising lots of money without much traction—in that sense, it’s still very much the pattern of, ‘I look like you, we have the same background, therefore I will back you,’ or they’re throwing around the right buzzwords, like AI. But that behavior is not happening with the frequency we saw before the bubble burst in 2022.”

Speaking of AI, Austin notes that including the term in a pitch deck has quickly become table stakes, although “if you’re seeing a distinct challenge or opportunity for certain business sectors due to generative AI, that can still be interesting.” In addition, how a business spends its funds is coming under even greater scrutiny. “Entrepreneurs are being asked for cash-flow hiring plans,” she says. “Telling a VC that you have a 20-person team when your venture is pre-revenue is not impressive; it could be worrisome.”

While the forecast is far from sunny, entrepreneurs are still raising funds, Austin notes—and plenty is still possible with little to no financial backing. “More founders are bootstrapping or taking smaller checks from angel investors,” she says. “Creating a business, and understanding the pain points, is still the greatest challenge for any new venture, regardless of the economic climate, but the barriers to entry have gone way down, thanks to some of the tools out there. You’d be amazed by what you can get done with very little money through generative AI and no-code tools, especially if you’re in software.”

Growing up with two parents who loved their careers as doctors, Simin Lee (MD 2015/MBA 2016) was drawn to their example. “I sensed early on that there was more to the business of delivering health care than making a diagnosis and prescribing medication,” she says. Today, as cofounder and CEO of Systole, a healthtech startup, Lee is developing a personalized, physician-driven digital exercise program for patients living with chronic conditions—such as heart disease and obesity—that bridges the gap between that high-octane spin class at a local gym and the joyless monotony of prescribed hospital rehab programs.

The startup is science-backed: Post-heart-attack patients who complete a monitored, 12-week program of walking on a treadmill or riding a recumbent bicycle can cut their risk of having another heart attack in half, says Lee, a cardiologist and behavioral scientist at Boston’s Brigham and Women’s Hospital. “In an era of so many incredible medications, that’s pretty impressive,” she notes. But the hospital-grade exercise isn’t usually much fun, so it doesn’t become a habit once insurance stops paying for the program; there isn’t the same social, sticky quality Lee herself discovered when she started rowing on the lightweight crew team in college.

“What if there was an option that took all the engaging, community-building aspects of the consumer-fitness experience and grounded it in medical science for a chronic disease population?” asks Lee, a 2023–2024 Blavatnik Fellow who will spend her fellowship year developing and testing an on-demand video prototype. Exercise can be an intimidating prospect for those with chronic conditions, but that can make a doctor-backed product like Systole a reassuring option, says Lee: “The patients I see are worried that the commercial options are too intense to be safe, or that they’ll be out of place if they aren’t a 20-something wearing Lululemon. There’s a big white space where we’re hoping to find a product-market fit with Systole.”

Generative AI’s capacity to transform business operations continues to be a hot topic, with a strong caveat in the form of threats to data security and privacy. Assistant Professor Seth Neel is principal investigator of the Trustworthy AI Lab at the Digital Data Design Institute at Harvard; his research develops tools for machine learning that mitigate bias and enhance privacy.

Generative AI poses a greater risk to privacy by its nature, Neel explains. A traditional machine-learning algorithm spits out a prediction based on data inputs, but it can be challenging to infer anything about the data from the result. “In a generative model, there’s a much greater capacity to repeatedly interact with and probe the model,” he says. “So it makes sense that this greater level of access potentially allows the extraction of more information from the underlying data. There’s a larger attack surface, if you will, for potential adversaries.” The models themselves also increase in size, he adds, with more parameters and a greater capacity to memorize data, creating ever-more sophisticated attacks. “It’s a perfect storm,” says Neel. “The question of exactly which points are memorized by the model, and why, is an area of research that is still pretty much unsolved and one of great interest to me.”

That answer also will be of interest to companies on the hook for potential security breaches, making a solution—or at least some sort of intermediate fix—even more pressing. “An open problem is whether you can train a generative AI model from scratch on your own data, preserve privacy, and still have good utility,” says Neel. Differential privacy—the practice of introducing a small amount of random “noise” to mask an individual’s data, while still ensuring the integrity and accuracy of the model, is one of many paths Neel is pursuing: “This will continue to be a tricky area for companies that want to harness generative AI tools based on customer data,” he says. “Given current systems, it’s not always clear how to deploy that information in a way that protects individual’s privacy. That’s where these privacy preserving techniques will have a huge potential impact, in my view.”

As Aetna’s CMO from 2016 to 2020, HBS senior lecturer David Edelman oversaw an initiative to encourage healthier and more cost-effective behaviors in the insurance company’s subscribers. The effort involved machine learning, with a team of engineers writing algorithms that spotted patterns in customer behavior—for example, a member with repeated ER visits—and generated a ranked list of possible prompts and desired actions (texting the location of a nearby urgent-care clinic as a less costly alternative).

Just a few years later, with the rise of AI, easily purchased (or free) AI tools exist to create an even more granular picture of customer behavior as well as many more testing scenarios, opening a world of opportunity to companies. To harness that power effectively, Edelman offers newbies some pointers on how to leverage AI’s shiny promise.

“It’s easy to be seduced by all of these cool technologies, but you need to think about AI integration along multiple dimensions,” Edelman says. First focus on clear, narrow goals that are more specific than “boost sales.” Next, consider your methods for collecting data; it should be both broad in terms of the range of data to capture and granular about attributes at the same time. Now, what will that data feed? An email management system? A forecasting model? A supply-chain management system? “One of the great things AI can do is help you better manage your business at a more micro level,” says Edelman. “If you have this AI engine, how will it change the way you operate? Whether that’s personalization for a customer, thinking through the right shelf assortment in a store, or stocking a warehouse, it implies changes in your process.” Finally, AI’s ability to rapidly test different options demands an experimental, supportive culture, requiring agile teams with the creativity to iterate over time.

While it’s still relatively early days in the AI revolution, Edelman advises against standing on the sidelines for much longer. “AI can accelerate the learning curve for improving a range of processes across functions, not just marketing,” he says. “Your competitors will be using it; if you don’t, you’ll be leaving value on the table.”

We’ve all experienced the staying power of a good story—the kind that gets told and retold at every family gathering. There’s good reason for this: The distinctive details and context of stories helps them stick in our memory in a way that numbers, which are more abstract, do not. As a behavioral economist with a focus on how information is communicated via natural language and how ideas become contagious, Assistant Professor Thomas Graeber has worked to understand the deeper reasons for this reality and its implications for anyone who hopes to make a lasting, convincing argument.

In a series of controlled experiments, Graeber and his colleagues studied how quickly different types of information dissipate over time, finding that the effect of a story faded by about 33 percent over the course of one day, whereas a statistic, by contrast, faded by 73 percent. “The temporal effect is much larger than we would have expected,” Graeber says. The most obvious takeaway is to employ a relatable, audience-specific story that ties into what your numbers indicate if you want to make your message stick.

But there’s another level to Graeber’s findings. Do you want the information you’re presenting to be acted on in the immediate future? If so, lean harder on numbers than stories. “Think about a politician out campaigning,” he says. “Is the election tomorrow or is it in a month? Statistics can more effectively inspire action in the short run. The further away you are from the point in time where the information will be used, the more you should lean into stories and anecdotes, because those are more likely to be remembered.”

“A big part of this study is trying to understand how memory really works,” explains Graeber. “There’s a unique quality to verbal communication and what people infer from different cues. What gets lost? What sticks? Understanding that has far-reaching implications in economics and business.”

In the dance between customer and brand, marketers often break down the experience into three moments: before, during, and after purchase. Research by Assistant Professor Julian De Freitas shows the advantage of considering the customer journey more holistically, and the proven benefits of ensuring that experiences stay consistently positive or improve over time, are punctuated by a moment of joy, and end with a “bang” of delight (think Disney’s end-of-day fireworks display).

The findings, while obvious, aren’t acted upon so easily. Getting a view into the customer experience is the first hurdle; De Freitas recommends taking a psychological, empathetic approach to how the customer might summarize his or her experience. This might involve going beyond star-type ratings to, say, enabling customers to continuously upvote or downvote the experience as it unfolds, to get a more nuanced picture of the customer journey. “If your marketing budget is stretched, you can get more bang for your buck by investing in certain aspects of the journey more than others,” De Freitas says. He advises maintaining a consistently positive or “building” experience: “The smoother and more positive, the better—avoid yo-yoing up and down.”

Finally, end on a high note: Positive final experiences have an outsized return on investment. “It’s easy to think that the customer experience ends when we have their money in our pocket, but as marketers, we focus on lifetime value,” says De Freitas. “The entire experience, not just the product, can be its own source of value.” Citing a family outing to Crate & Barrel, De Freitas describes the care a sales associate showed in packaging their purchases and loading them into the car. It could have ended with the painful process of lugging and juggling, all while balancing the needs of small children; instead, the pain-free experience offered just the sort of fireworks effect that leaves a lasting afterglow.

Money will always have its place as an employee motivator. But other levers can be equally—if not more—powerful, says Professor Brian Hall.

If mission and values are aligned, a company’s culture alone can be highly motivating, notes Hall—and that’s where it all begins: “When people ask me, ‘Does the compensation system determine the culture, or does the culture determine the compensation system?’ I say, ‘Yep.’ They’re tied to one another. What you incentivize tells a story.” For that reason, he adds, it’s a good idea to include values like teamwork or customer relations in incentive plans that are difficult to measure but key to a company’s success.

That goes for non-monetary awards as well. Dig deep to consider whether that Employee of the Month plaque has real meaning that ties into your company’s purpose. And carefully consider the benefits you offer. Hall offers the example of Wegmans, which began an effort in 1984 to attract hardworking teenagers interested in advancing at the supermarket chain. They introduced a partial scholarship program, which in turn drew a pool of motivated, growth-focused applicants. “Some of these decisions may seem like an afterthought,” Hall observes, “when in fact they can be extremely strategic.”

Finally, as robotics and AI sub in for various human tasks, Hall asks employers to go beyond the obvious cost-saving benefits to determine technology’s impact on workers. In other words, is the innovation making the job more, or less, engaging for employees? “You don’t want to automate the part of work that people really enjoy,” he says. “People want to use their minds. Autonomy in and of itself is a reward.”

Imagine generating $1.7 billion in sales in a 12-hour live stream. That record-setting feat, achieved on the Chinese platform Taobao, belongs to Austin Jiaqi Li, an online influencer with an estimated net worth of more than $15 million. Seeking to better understand this powerful phenomenon, Assistant Professor Jeremy Yang and his collaborators developed an algorithm by evaluating 40,000 influencer ads on TikTok and derived a product-engagement score for each video ad to predict the effect on product sales. Yang, who develops data products for advertising, targeting, and pricing decisions in his research, is more broadly interested in the “creator economy” ecosystem that includes influencers like Li; platforms like TikTok, Instagram, and YouTube; and the brands whose advertising dollars fund most of this activity.

To start, Yang notes that not all companies should sell through social media. Does your customer base fit the demographics of a platform like TikTok? Is your product suited to the medium? “It’s probably more effective to sell cosmetics on TikTok than retirement financial products,” he says. Consumers also behave differently across platforms: “I go to TikTok to watch dance videos with no clear purchasing intention; but I might see an ad for something that grabs my attention, which is a very different dynamic than a search-driven purchase on a site like Amazon. For that reason, you need to engage the user first, rather than selling upfront.” This requires a platform-native approach, says Yang: a video that works well on Amazon would appear as a hard-sell turnoff on TikTok.

Working with an influencer also requires forethought—and some caution. “It’s important to remember that customer loyalty will be to the influencer first—not necessarily your brand—just as we go to Walmart to make purchases without a specific brand in mind. Influencers are becoming the new Walmarts of the world,” says Yang.

Post a Comment