THE SOLD OUT EFFECT: How Aesthetic Appeal and Limited Edition Status Shape Future Purchase Intentions Across Generations
Advisor: Lew Gleich
Internship Location: Florelle Inc.
Onsite Mentor: Huaqing (Emma) Liu
When products sell out, the conversation around them often intensifies rather than disappears. "SOLD OUT" posts on TikTok and Instagram regularly generate more engagement than "available now" content. During my internship in the beauty and skincare industry, I will observe this phenomenon firsthand, examining how brands use unavailability as a promotional tool. Yet consumer psychology research almost entirely overlooks what happens to purchase desire after a product becomes unobtainable. I’ll investigate how aesthetic appeal and limited-edition status interact to shape future purchase intentions once a product has sold out, and whether age moderates this effect. My findings will offer brands concrete, evidence-based guidance on whether publicizing a sellout is a strategy worth pursuing.
Project Blog
Post 1: Introduction
Welcome to my first Senior Project blog post! My name is Jenna, and I'm a senior at BASIS Independent McLean. For my Senior Project, I'll be working with Florelle Inc., a growing beauty e-commerce company based in Lorton, Virginia, that sells skincare and beauty products across multiple online marketplaces.
My interest in this work comes from prior experience at a furniture company, where I worked on social media campaigns and noticed how consumer engagement shifts depending on how and where products are presented online. I've always been curious about how brands connect with consumers, especially in digital spaces, and this internship felt like a meaningful opportunity to explore that further.
My role at Florelle Inc. will focus on digital marketing and e-commerce operations, which includes live streaming, content creation, video editing, managing social media accounts, and analyzing platform analytics to support their growth in the U.S. market.
Before officially starting on-site, I've been getting a head start on the research side of things, right now I'm working on creating my pre-tests and general surveys, while also looking into platforms to distribute them through. These will help me build a clearer picture of consumer behavior and digital marketing engagement going into the internship.
I'll be updating this blog weekly with what I'm learning and how the project develops. Thanks for reading!
Week 1: First Week In
Welcome back to my blog! This past week marked the official start of my on-site internship at Florelle Inc., and it ended up being a much busier week than I expected.
On my first day, my on-site supervisor walked me through the basics of the company that was quite new to me—Florelle is a growing e-commerce business that sells across multiple platforms, primarily focusing on Whatnot and Temu for now, while also selling full-size products in Hong Kong. Their U.S. inventory focuses heavily on sample-size fragrances from well-known brands like YSL, Tom Ford, and Dior. It's actually a pretty interesting business model, because I hadn't really considered before how much a continuous supply chain between two markets changes the way a company can operate, and working with sample-size luxury products turned out to be more involved than I initially assumed. After getting familiar with the product range, I spent the rest of the day taking product photos, for example: cover images and detail shots for the selling pages and started learning the basics of Whatnot, including how to upload products, navigate their Excel sheets, and work with barcodes and SKUs.
The second day was more of the same, but this time we switched to a camera for the product photography. I also started getting into the U.S. marketing side of things, where my supervisor walked me through what tends to perform well with American customers, what kinds of products they're more likely to engage with, and how to approach pricing accordingly. It was interesting to see these decisions play out in a real business context, especially since a lot of what drives consumer engagement here — brand recognition, product presentation, perceived value — ties directly into what I've been exploring in my research on consumer psychology and purchase behavior.
By the third day, I had a much better handle on the uploading process for Whatnot and things moved a lot faster. I also started learning about Temu, which has a pretty different workflow. It involves formatting things through Excel first, which takes more time but keeps things organized. On top of that, I worked on making product cover images in Canva, which felt familiar since I had prior experience with it. To wrap up the week, I'll be having my first live stream on Whatnot on Friday, where I'll be advertising our products and getting a feel for the preparation that goes into it.
On the research side, I've been continuing to develop my pre-tests and general surveys. Most of this week went toward the internship, but I've been making gradual progress such as refining the questions, thinking through what I actually want to measure in terms of consumer behavior and digital marketing engagement, and continuing to look into the best platforms for distributing them.
Thanks for reading,
Jenna
Week 2: Finding the Flow
Welcome back! This week has been dominated by live streaming on Whatnot, while also continuing to work on the Temu side of things in parallel.
The live streaming experience on Whatnot has been interesting. Whatnot operates as a bidding platform, so viewers bid on top of each other to win items. It's a pretty different dynamic from regular e-commerce, and it took some getting used to. The first stream didn't go too smoothly, but each show since then has felt more natural as I've gotten a better feel for the pace and flow.
One of the more interesting things I've noticed through streaming is how much brand recognition and product design actually influence bidding behavior in real time. Since we're reselling sample-size fragrances from well-known brands, it became pretty clear early on that certain brands perform significantly better than others. Kayali, for example, turned out to be really popular with U.S. audiences, which I hadn't expected, especially coming from a background where that brand wasn't particularly well-known. Meanwhile, brands like YSL attracted buyers more through packaging and design appeal than brand loyalty alone. It's a good reminder that what sells in one market doesn't always translate to another, and that consumer behavior is shaped by a mix of cultural familiarity, aesthetics, and perceived value. I also noticed that mentioning scarcity, like letting the audience know there were only one or two left in stock, consistently drove up bidding. That connects pretty directly to what I've been researching around limited availability and purchase behavior.
Another thing we've been figuring out as we go is product selection before each show. Since Whatnot is essentially a wholesale clearance platform for us, selling at or sometimes below market price to move inventory, we've been experimenting with how to group items. At first we tried sorting by brand in sets of five or ten, and we even opened up Sephora bundles to regroup similar items together. In practice though, we ended up pulling products more spontaneously during the show anyway, so a lot of that prep didn't get used the way we planned. It's still something we're working out.
We also made a change to how we handle pricing during the show. Instead of adjusting prices mid-stream, we now start everything at $1 and let the bidding run. It keeps the show moving faster and makes the post-show order organization a lot cleaner since everything is numbered in sequence.
The post-show shipping process has honestly been the most demanding part. Each stream can move anywhere from 200 to 300 items, and since we stream from the office while some inventory is stored at a separate warehouse, there are times when items need to be retrieved later. We label everything individually to keep track, but with that volume it's easy for one or two things to end up in the wrong package, which means going back through everything to fix it. By the end of each show I'm pretty exhausted. Between handling product selection, running the stream, and working through shipping, it takes a lot out of you, and my throat has been feeling it too. But I've noticed myself getting faster and more confident with each session, which makes it worth it.
On the research side, I've made some progress this week. I set up my pre-test survey on Qualtrics, which I chose over Google Forms because it allows for random assignment to conditions, something that's important for the experimental design of my study. I'm currently working out how to distribute it, and I'm planning to use Prolific, a participant recruitment platform that lets me filter by demographics. That same approach will likely carry over to my full survey as well.
Until next time,
Jenna
Week 3: The Grind
Welcome back! This week felt like a natural continuation of the first two, but with a lot more on my plate and a lot more independence in handling it.
The biggest highlight of the week was breaking our sales record, bringing in around $3,000 worth of products in a single stream. For context, we're selling sample sizes, so that number is genuinely impressive. I think a big part of it came down to consistency, because by this point we have a fixed streaming schedule, so returning customers are actually bookmarking our shows and showing up at the start. We also hit over 600 followers by the third week, and I noticed an unusual spike in new followers right at the beginning of Thursday's show, which felt like a sign that things are starting to gain traction. For that particular stream, we shifted toward a wholesale model, selling bundles of around 10 vials at a time rather than one or two pieces, and I think that approach worked well. We opened with a Kayali bundle to draw people in, and it seemed to set the right tone for the rest of the show.
One moment that stuck with me this week was having to handle audience questions on my own before my advisor arrived. People were asking about pricing, and I didn't always know what the company was comfortable selling certain products at. I made judgment calls and it worked out, and my advisor was actually pretty relaxed about it afterward. She mentioned that losing money is sometimes a necessary part of building a market presence, which was a genuinely useful perspective to hear at this stage.
On a typical streaming day, I start at around 10 AM and go until the early afternoon, take a lunch break while my advisor covers the stream briefly, then come back for another stretch. After the stream wraps, the packaging and shipping process begins, and that's where the long days come from. Some days I've stayed past 6:30 working through it on my own — finding products, matching them to orders, labeling everything so items from the warehouse can be tracked down later. With 200 to 300 items moving per show, even small mix-ups take time to fix. It's physically and mentally tiring, but honestly I prefer the physical exhaustion. The days where I'm sitting at a computer uploading products or editing photos for hours feel draining in a different way.
Speaking of which, this week I also started relearning Photoshop to work on product cover images. Since customers shopping online can't physically see or smell what they're buying, the front page photo does a lot of heavy lifting: it's often the only thing that catches someone's attention before they decide whether to keep scrolling. Creating an effective template for fragrance samples has been more challenging than I expected, partly because my Photoshop background is mostly in portrait photography, so product editing is a different skill set I'm still building. I tried several template versions this week but haven't landed on one that fully works yet.
I also spent some time this week watching how other streamers run their shows to pick up techniques. One thing I noticed is that sellers doing wholesale tend to bundle sample vials in groups of six or more, which moves inventory faster. The reasoning my advisor gave is that when prices are bundled, customers are less likely to mentally compare it to the individual market price, which actually makes them more willing to bid. It's a subtle thing, but it's a good example of how consumer psychology plays out in a very immediate, visible way during a live stream — something that keeps feeding back into how I'm thinking about my own research.
On that note, I didn't make much progress on the research side this week. The internship days have been long enough that by the time I'm done, there isn't much left in the tank. I'll be picking that back up next week.
Thanks for reading,
Jenna
Week 4: Lessons from Streaming and Literature
Welcome back! This week felt like more of the same on the surface, but there were some meaningful changes underneath.
On the streaming side, things ran noticeably smoother this week thanks to a better system for tracking our inventory during shows. Instead of trying to organize items on the fly while streaming, we now scan barcodes directly into a spreadsheet where an AI-integrated tool automatically pulls up the product name, category, and stock quantity. It sounds simple, but it made a real difference in how quickly the show could move. That said, I did run into some problems. I lost track of how much stock we had for certain items and ended up overselling, which meant my advisor had to call me mid-show to tell me to pull those products. A little embarrassing, but a good lesson in keeping closer track of inventory before and during a show.
The bigger new addition this week was customer service, which turned out to be more interesting than I expected. A large portion of the refund requests we received were from customers complaining that their fragrance samples were too small, even though these are items that have "mini size" and "travel size" clearly labeled in the listing title and images, with the exact volume printed on the product itself. It was hard to understand at first, and honestly a little funny. But it also got me thinking about something more serious: if customers consistently overlook product details even when they are clearly displayed, that says something about how people actually process information when shopping online. It lines up with what research on consumer decision making suggests, which is that people often rely on emotional or intuitive cues rather than carefully reading the details in front of them, especially in fast-moving digital environments. That is part of what makes product presentation so critical in e-commerce, and something I have been thinking about more as I work through my own research.
I also started communicating more directly with the warehouse team this week, mostly through group messages. It is something I did not really anticipate — working with people who have a lot more professional experience than me has made it harder to communicate effectively, and there have been moments where my messages got overlooked until my advisor followed up. It is a work in progress, but I think it is a good thing to figure out sooner rather than later.
On the research side, I spent some time this week reading through a few papers to refine my literature review. One that stood out was Huijsmans et al.(2019), which used fMRI imaging to show that a scarcity mindset actually changes the neural processes involved in consumer decision making, specifically increasing activity in the orbitofrontal cortex, a region linked to valuation, while reducing activity in the prefrontal cortex, which is associated with goal-directed reasoning. That is a more biological angle than I expected to come across, but it adds weight to the broader idea that scarcity does not just influence what people buy — it actually changes the way they think. I also read Chang and Chang(2023), which looked at how digital media and social networks shape Gen Z's attitudes toward unfamiliar products, and found that digital media has a significantly stronger influence on Gen Z than on older generations. That is relevant to my research given that a lot of what I am looking at involves how aesthetic presentation and scarcity signals affect purchase intentions in digital spaces. And Przybylski et al.(2013) on Fear of Missing Out added another layer to my literature review. The idea is that social media engagement is partly driven by a psychological need to stay connected and not miss rewarding experiences, which connects to how limited availability messaging works in online selling.
Until next time,
Jenna
Weeks 5-6: Survey Progress and Lots of Streaming
Welcome back! Since I worked through spring break instead of taking time off, this post covers two weeks rather than one. Weeks 5 and 6 have been pretty similar to the previous ones, so I will keep this one relatively brief and focus on what is actually new.
The biggest shift is that I am now streaming every single day at this point, fully on my own. It is tiring, especially on my throat, but honestly I find it a lot more engaging than sitting at a desk doing customer service or editing spreadsheets for hours. One thing that has made it more rewarding recently is the kind of feedback I have been getting from viewers. People have been leaving really warm messages, thanking me for offering good prices, saying they appreciate how I run the show, and mentioning that I actually listen to their requests during the stream. That kind of direct feedback is not something you get from most tasks, and it has made the daily grind feel more worthwhile. The account is also approaching 1,000 followers, which feels like a significant milestone considering the account only had around 90 followers when I first started in early March.
Customer service has taken up a lot more of my time these past two weeks. I check and respond to messages twice a day, once in the morning and once at the end of the day. A big part of the job has been requesting proof videos from the warehouse for specific orders, which has been more frustrating than expected. One of the warehouses is managed through an outsourcing company, and getting a response from them usually requires asking twice, sometimes across two days, just to get a video for a single tracking number. Sending a full list does not help either since they will only respond to requests one at a time if you follow up individually on each one. It is a slow process and something I am still figuring out how to navigate without constantly having to loop in my advisor.
On top of that, since most of my coworkers handling the Temu platform are based in China and the seller portal requires a U.S. login, a lot of those responsibilities have been shifting to me. There is a backlog of product listings that need additional information added, things like net weight and other details that the platform has started requiring more strictly, and I have been working through those alongside everything else. It has been a lot to keep up with.
On the research side, things are actually moving forward in a concrete way. I posted my pre-test survey to Prolific earlier this week and am waiting on around 20 responses, which I expect to have back by this weekend. Prolific charges to recruit participants, but being able to filter by demographics makes it worth it for what I am trying to measure. In the meantime, I have been working on building out my main survey on Survio, which is a new platform for me. It is structured quite differently from Google Forms, more like building a questionnaire layout from scratch rather than a straightforward form, so there has been a learning curve in figuring out where everything is. I am hoping to have the main survey closer to finalized once I see how the pre-test results come back.
Thanks for reading,
Jenna
Week 7: First Survey Results
Welcome back to my blog! Not too much has changed on the internship side this week, but there is one notable shift: we moved our streaming setup from the office to the warehouse, which means I now have direct access to a much larger range of products during shows. In the time between streams, I have also been helping sort and classify incoming stock since new products arrive daily, and there are only two people managing that full time. On a more practical note, the warehouse is about five minutes from my house, which is a really welcome change from the 20 to 30 minute drive to Tysons Corner.
On the research side, I got back my pre-test results this week, which was exciting. The survey had 21 responses in total, and the results were pretty much what I was hoping for. Most participants were familiar with reusable water bottles and tumblers and said visual design mattered to them when evaluating products. When it came to rating the two images, Product Image A came out on top across the board — it scored higher on appeal, attractiveness, design, and how premium it looked, while Image B scored lower across those same measures. 18 out of 21 people picked Image A as the more aesthetically appealing one overall, with 2 picking Image B and 1 saying there was no real difference. In the open-ended responses, people generally described Image A as clean and minimalist, while Image B was seen as too busy or overwhelming. Overall I think the pre-test confirmed that the two images are being perceived the way I intended, which means I can move forward with the main survey feeling a lot more confident about my stimuli.
Speaking of the main survey, I have already sent out half of the eight survey versions and am waiting on those results before sending the rest out this weekend. Since my study uses a 2x2x2 experimental design, I ended up making eight separate survey versions, each with a different variation of the product exposure page. I built the product pages using Canva and Shopify to create realistic-looking mock listings. For the survey platform, I went through a bit of a process: I tried Survio first, then SurveyMonkey, but the features I needed were not quite worth what they were asking for, so I ended up going back to Google Forms, which is straightforward and works well for what this study actually requires.
I am planning to launch the main survey soon and will share updates on the data next week!
Week 8: Main Survey Results
Welcome back to my week 8 blog post. I will be honest, this was probably the most tiring week of the internship so far, and I say that having gone through some already pretty long days.
The workload has ramped up a lot recently. A typical day now looks something like this: customer service in the morning, streaming from around 10 to 1, a very short lunch break that rarely feels like a full break, then back to packing and preparing the shipping list for the morning show, picking and scanning items, and then if there is time left, counting stock in the warehouse to make sure numbers match what is on the box. On top of that, we have added an afternoon show, which adds another full round of everything streaming related to the day. It is a lot to manage at once, and I will admit it has been hard to keep up with at the pace expected. That said, I think learning to handle that kind of volume and pressure is probably one of the more useful things I will take away from this experience.
On a more positive note, I have started to get a clearer read on what affects our show performance from week to week. Things like how many viewers the platform promotes us to on a given day, whether we have competition streaming at the same time, and even just whether the products we pull that day happen to match what people are in the mood to buy — all of these factor in more than I initially realized. It has been an interesting thing to observe in real time, and it connects more than I expected to some of the ideas I have been exploring in my research around how availability and presentation influence consumer interest.
Speaking of research, the main survey results are in, and I finally have something real to look at. I ended up with just above 140 responses spread across all eight conditions, which matches what I was aiming for, but it was not cheap to get there, because recruiting through Prolific for all eight surveys ended up costing me around $100, which was a little bit more than I expected, but I think the quality of the responses makes it worth it.
Looking at the data at a surface level, a few things stood out right away. First, the high aesthetic conditions consistently rated higher on aesthetic appeal than the low aesthetic ones, which is expected and confirms the manipulation is working. What was more interesting was looking at future purchase intent across conditions. The highest purchase intent actually showed up in the high aesthetic + regular + sold out condition, which scored noticeably higher than most other groups. The low aesthetic + limited edition + sold out condition also came in with a relatively high purchase intent score, which was a little surprising given that the product was rated lower on appeal. On the exclusivity and brand perception side, the limited edition + sold out condition under high aesthetics scored the highest on brand status perception, while the low aesthetic conditions generally came in lower across quality and worth its price ratings. The "missed opportunity" and "want updates on future launches" scores were relatively low across most conditions, which is something I want to look into more carefully. I am not drawing any conclusions yet since I still need to run the actual statistics, but having the data in hand after all the time spent building and distributing the surveys is a really good feeling.
Until next time,
Jenna
Week 9
Welcome back! This is the second to last week of my internship, and things are still moving at a pretty fast pace even as the end approaches.
The biggest new thing this week on the internship side is that my advisor wants to start doing TikTok promotions. Since I will be wrapping up soon, I was mainly helping get things set up for after I leave. These will include building the account, filming product showcase videos, and editing them for posting. The style was mostly aesthetic product clips with text overlays and music, similar to what you would normally see on TikTok for this kind of thing. My advisor had a general direction she wanted, but gave me some freedom with the music choices, which I appreciated. Honestly, this part felt the most natural to me out of everything I have done here, and it was a nice change of pace from the more repetitive parts of the job.
Everything else this week was pretty much the same as before, from streaming to customer service to warehouse work, the usual. I will save the bigger reflections for next week's post.
On the research side, I have been deep in the final draft of my analysis, and there is one finding I want to highlight here that genuinely surprised me. When looking at how different age groups responded to the sold-out versus available conditions, Gen X participants — the oldest group in my usable sample — actually showed higher purchase intentions when the product was available compared to when it was sold out. This is the opposite of what the literature and my hypothesis predicted. The expectation was that sold-out status would drive up interest and desire across the board, but for this age group, it seemed to work the other way, with availability reading as a more positive signal than scarcity. Gen Z and Millennials followed a more expected direction, though none of these differences reached statistical significance. It is one of those findings that makes you want to dig deeper, and it is something I plan to discuss more fully in the paper.
I am also still working on the side experiment, which is a hypothetical $200 consumer choice exercise where participants select items from a basket of available and sold-out products. The main challenge right now is figuring out how to make the price automatically calculate as participants add items, the way a real shopping cart would, since I do not want them to have to do the math themselves. I am considering whether a mock Shopify storefront might work better than Google Forms for this, though I am still figuring out how to incorporate the waitlist option for sold-out items since Shopify does not support that natively. Hoping to have that sorted and sent out soon.
Thanks for reading,
Jenna
Week 10: The Final Week
Welcome back for the last time — at least for this internship. This week officially marked the end of my time at Florelle Inc., and it looked a little different from the rest. Instead of streaming, I spent most of the week on handover work: finishing up product photography for items that still needed proper shots, and spending two full days with four people recounting all of our warehouse stock to make sure the numbers were accurate before I leave. It was genuinely one of the more tedious things I have done here, but I would rather end on something that actually needed to get done than nothing at all.
Looking back across all ten weeks, the thing that surprised me most was how much I actually enjoyed this. Going into it, I did not have a strong picture of what working in e-commerce operations would feel like day to day, and I did not expect to find the work as engaging as I did. Working nonstop from 9 to 6 sounds exhausting on paper, and some days it really was, but I rarely dreaded coming in the next morning, which honestly surprised me.
The growth I am most proud of is not any one thing; it's all of the things together. Going from not knowing how a live stream auction works, to running daily shows completely on my own, growing the account from around 90 followers to almost 1,500, learning how to handle customer complaints, coordinating with the warehouse, figuring out product photography, Canva, Photoshop editing, and just getting faster and more confident at everything with each passing week. None of it happened at once, and I think the hardest weeks were actually the ones that moved things forward the most. I also genuinely appreciate that this opportunity came in an industry I actually care about. I have always been interested in beauty and skincare, and getting hands-on experience inside a company that operates in that space, even at the sample and e-commerce level, feels like a real starting point. If this is a direction I want to continue in, I now have a much more concrete sense of what it actually looks like from the inside, which I did not have before.
On the research side, the side experiment using Tally is just about ready to go out. Tally ended up being the right platform for this, compared to Google Forms and Shopify; it handles automatic price calculation as participants select items, which was the main thing I needed that Google Forms could not do, and Shopify would have overcomplicated. I am planning to send it out this weekend and wrap up that piece of the project shortly after.
Working here for ten weeks also gave me a more grounded understanding of what I was actually studying. Watching customers respond in real time to scarcity cues during live streams, noticing how mentioning limited stock would immediately drive bidding up, seeing which brands generated more interest than others — all of that gave me a more concrete feel for the consumer psychology behind my research than I had when I started. The literature describes these patterns, but seeing them play out in front of you, in real time, during a live auction, makes the ideas feel a lot less abstract.
Thanks for reading,
Jenna
Week 11: Wrapping Up
Welcome back! This one will be a shorter post since the internship officially wrapped up last week, and most of my time this week has gone toward finalizing my paper draft and putting together my presentation slideshow.
The main thing I can share is that the results from my side experiment are in. The Shopping Choice Exercise was completed by 22 participants across three age groups — Gen Z, Millennials, and Baby Boomers — each given a hypothetical $250 budget to spend across a catalog of 30 products, three of which were sold out with a waitlist option. Overall spending was pretty consistent across groups, averaging around $227 out of $250, which tells me that budget allocation behavior did not really differ by age. The more interesting measure was waitlist joins, which were the main behavioral thing I was looking at. Out of 22 participants, only 3 joined any waitlist at all: two Baby Boomers who chose the floral tote, and one Gen Z participant who chose the gaming mouse pad. The most expensive sold-out item received no joins from anyone.
What stood out to me was that the people who did join waitlists seemed to be drawn to items that matched their personal taste rather than just responding to the sold-out label itself. The Baby Boomers went for the aesthetically distinctive tote, and the Gen Z participant went for the gaming item. That suggests aesthetic relevance to your own interests might matter more than scarcity framing on its own, which is something I want to discuss in my paper. The overall low waitlist rate also lines up with the null results from the main survey, where sold-out status did not produce significant increases in purchase intention either. Both point in the same direction, and figuring out why is something I am leaving for the conclusion.
Thanks for reading,
Jenna