Research, in Plain Language

For a long time, firms have relied on opinion leaders — individuals who influence others by sharing their experiences — to market their products. With the rapid growth of social media platforms such as Facebook, Instagram, LinkedIn, and X, firms have turned to influencers on these platforms to reach customers. A survey by the Influencer Marketing Hub (2022) reports that the market size of influencer marketing grew from $1.7 billion in 2016 to nearly $13.8 billion in 2022, with almost 75% of respondents planning to run influencer campaigns.

With thousands of influencers to choose from, a key question is: given a limited budget, how should a firm design its influencer marketing campaign? Designing a successful campaign entails selecting an optimal subset of influencers to promote the firm's products — and, when campaigns run over longer periods, scheduling those influencers' ads over time.

Our Management Science paper, "A Framework for Analyzing Influencer Marketing in Social Networks: Selection and Scheduling of Influencers," develops a data-driven modeling framework to help firms conduct both short-horizon and long-horizon campaigns. The models are grounded in interactions with marketers, observation of firms' message placements, and parameters estimated empirically using data from Twitter.

Our empirical analysis uncovers the effects of the collective influence of multiple influencers and identifies two critical parameters: the multiple exposure effect — the curvilinear effect of additional ad exposure on a customer's engagement — and the forgetting effect — the impact of the gap between two consecutive ads on engagement, an important factor when designing long campaigns.

For short-horizon campaigns, we develop an optimization model to select influencers. While the problem is computationally hard, we use a mathematical programming procedure to provide near-optimal solutions. The paper also demonstrates the economic significance of the peer effect — the influence of one follower on another — which is particularly significant in networks with many connections among followers. Notably, selecting influencers based on their follower counts alone is not always the best strategy.

Long-horizon campaigns (appropriate for brand building) require careful selection of influencers plus a schedule that sequences ads over weeks or months. We develop an efficient procedure that simultaneously selects influencers and schedules their posts. Further analysis shows that the firm's net benefit is concave in the budget — it first increases at a diminishing rate and then declines — highlighting the need to set campaign budgets strategically. We also show that random scheduling can lead to sub-optimal campaign performance.

Customers expect fast, on-time deliveries when they shop online, yet delays are inevitable — especially during demand surges such as the holiday season. Most customers have grown used to tracking their orders on the seller's website while the order is still being processed. Major retailers such as Amazon and Walmart have invested heavily in these technologies, and smaller online retailers can now do the same through ready-to-use, cloud-based solutions such as ClickPost and LateShipment.

Our study in the Journal of Operations Management examines how giving customers the ability to track orders affects satisfaction. Using data from a major online retailer, we find that on-time delivery is no longer the only thing that matters once customers can track their orders. Even when orders arrive on time, customers can feel dissatisfied if the retailer took too long to process the order before handing it to the logistics partner (3PL). This negative effect is more severe than a slow 3PL delivery, because customers blame the retailer for failing to process orders promptly.

Allowing customers to track order status can therefore expose deficiencies in a retailer's internal fulfillment process — regardless of whether the order was ultimately delivered late or early. While timely delivery clearly matters, customers also expect to see the retailer's effort in promptly processing their orders.

The study offers several recommendations. First, retailers that allow order tracking should prioritize improving internal operations (picking and packaging) so orders move quickly. They can also ask customers to review the product and the delivery separately, so fulfillment delays are less reflected in product reviews.

Second, and more importantly, customers may feel dissatisfied the moment they see a delay in the status page — even without knowing whether the order will actually arrive late. Borrowing from service-recovery research, retailers should proactively reach out when processing time runs long. A simple message — "there is a slight delay in processing your order, but you can still expect delivery on time" — can go a long way. Finally, because the positive impact of early delivery is minimal compared with the negative impact of late delivery, retailers should under-promise on processing time to avoid surfacing internal delays.

By taking these steps, online retailers can improve customer satisfaction and better meet expectations around delivery.