Cracking the Cold Start Problem
In this episode of Data Skeptic, we dive deep into the technical foundations of building modern recommender systems. Unlike traditional machine learning classification problems where you can simply apply XGBoost to tabular data, recommender systems require sophisticated hybrid approaches that combine multiple techniques. Our guest, Boya Xu, an assistant professor of marketing at Virginia Tech, walks us through a cutting-edge method that integrates three key components: collaborative filtering for dimensionality reduction, embeddings to represent users and items in latent space, and bandit learning to balance exploration and exploitation when deploying new recommendations. Boya shares insights from her research on how recommender systems impact both consumers and content creators across e-commerce and social media platforms. We explore critical challenges like the cold start problem—how to make good recommendations for brand new users—and discuss how her approach uses demographic information to create informative priors that accelerate learning. The conversation also touches on algorithmic fairness, revealing how her method reduces bias between majority and minority (niche preference) users by incorporating active learning through bandit algorithms. Whether you're interested in the mathematics of recommendation engines or the broader implications for digital platforms, this episode offers a comprehensive look at the state-of-the-art in recommender system design.
Guest
Boya Xu: Boya Xu is an Assistant Professor of Marketing at the Pamplin College of Business, Virginia Tech. Boya holds a Ph.D. in Marketing and an M.A. in Economics from Duke University, and a B.S. in Statistics from Zhejiang University, China. Boya conducts research in quantitative marketing, focusing on platform designs, content marketing, and the economies of emerging technologies. Her research combines methods from econometrics, online experimentation, and unstructured data analytics. Boya was the winner of the 2024 Doctoral Dissertation Research Award at the American Statistics Association (Marketing Section) and a recipient of the 2023 NET Institute Research Grant.
Transcript
Kyle Polich : Welcome to Data Skeptic a podcast exploring the methods use cases and consequences of recommender systems
Kyle Polich : Welcome to another installment of Data skeptic Recomer Systems
Kyle Polich : Today we get into the real nuts and bolts of what it is to build and design a recommender system You know in machine learning when you look at classification everything is very standardized Get your data into a nice tabular format one objective column pick an algorithm probably XG boost see if the model can learn to predict the objective from the features available But you can't just graph that model onto recommender systems
Kyle Polich : In fact most pathways to success involve a hybrid model
Kyle Polich : Today we talk about an approach that's a hybrid method combining several components collaborative filtering the classic recommender system algorithm which effectively does dimensionality reductions Then we talked about the use of embeddings how you can recommend your users and their attributes and things like that in a latent space And last but not least bandit learning Unlike some classification problem where you have presumably a well sampled data set
Kyle Polich : You never know how a user is going to react to a recommender system change
Kyle Polich : When you deploy a new idea you've got to explore it and when it works you've got to exploit it Anyway that's just a quick overview of some of the interesting topics we're gonna get into right now
Boya : Hello I'm Boya I'm from Virginia Tech I'm an assistant professor of marketing in the Pomplin College of Business Before joining Virginia Tech I had my PhD study in marketing from Duke So I actually graduated last year And before entering the PhD program in marketing I had a two year study of economics as a master's student in Duke as well Before that I had my undergraduate in statistics from China
Kyle Polich : Could you talk a little bit about the transition I mean you didn't leave academia but you went from being a student to being in the current position you're in now Could you talk a little bit about that transition
Boya : So I think that the biggest difference that I can tell from this transition is your role because that when you were a student you just focus on your projects And also in Duke that PhD students generally don't need to teach because we don't have
Boya : The undergraduate program for like PhD students to teach in a school But so far because you have become an assistant professor formally so you need to make some balance between teaching and research So definitely you should be a good time uh planner
Boya : And also I'm not only working on my own projects but I'm also collaborating with the students and try to help them to figure out their research interests and also figure out our interests in common and then we can develop some projects and also I will provide some advice to the students So that's something different But in general the research routine doesn't change that much because I I still work on some projects that I started during PhD but haven't finished before the graduation
Kyle Polich : And can you share some details on your research interest and the things you're currently looking into
Boya : My research interest is mainly about the digital economy especially the digital platform and I focus both on the e-commerce platform and also how the technology such as the recommendation system could help consumers and also how it would affect the supply side that is the uh the service or content provider
Boya : In addition to the e-commerce platform that another stream of my research focuses on social media I care about the content strategy on social media such as dynamics of the influencers production of controversial content and also how the news circulated on social media affects the offline business outcomes such as how the news about Impossible Meat affects the local restaurants adoption of this new food technology
Boya : So that is the overall of my research and to handle all the questions I just mentioned that I usually use the combination of tools including the econometrics causal inference and machine learning tools
Kyle Polich : When you study e-commerce specifically uh as far as I know pretty much e-commerce stores are like closed private businesses and they're not necessarily obliged to share any data Are there challenges or maybe on the positive side what data sets do you have available to study e-commerce
Boya : Yeah that's a really good question because you actually figure out the the main challenges that many researchers especially in the university are facing It's pretty hard to conduct a field experiment with the communists because they have so many other priorities for their KPI And also if you want to collaborate with them for maybe such as conducting a field experiment that you usually need to collaborate with multiple departments because that even for the
Boya : Development of recommendation systems that the departments in the firm that could be involved would include the uh search engineering team and also the product manager as well as the team for AB testing and so forth So it's usually pretty hard unless that you would have some internal contact Otherwise that we usually rely on some public data sets such as the archive data shared on uh Haggle to understand some historical
Boya : consumer behaviors Or another way is to establish your own platform for research such as the research platform used in my paper for recommendation system was established by a group of behavior scholars from Duke that they mimic the real grocery shopping scenario and we can recruit participants to join our experiment And
Kyle Polich : at
Kyle Polich : what point do you first take an interest in recommender systems
Boya : The the main reason why I started the research is from my own experience because you know the e-commerce is almost everywhere
Boya : So for me even if some purchases were still happen offline many of them actually happen online right And every time when you browse an online store that you realize the store recommends you to recommend something to you sometimes the recommendations are good sometimes are bad And another thing is that I think my search the entire shopping journey was affected by the recommendations
Boya : Because it affects what products that I can get exposure Sometimes that my next search is usually based on the previous search or the previous uh browsing right But the previous browsing sometimes was recommended by the website itself So that is the one scenario The other scenario is not limited to the e-commerce setup it's more about the the content consumption For example when you
Boya : You uh browse the TikTok or browse the Facebook or Instagram that you can always see some recommended content and also some content are the organic content which means that they're just for entertainment But some are actually apps especially Instagram I find that the recommendation there could really understand what I would like Many times that I was pretty attracted by the
Boya : Or some furniture recommended on Instagram and then I was directed to the e-commerce platform through the recommendation on social media So you can tell that how big influence it has at least to my own life And it's the same case for my friend's life So I can tell that the recommendation assistance is actually affecting people's life in so many ways
Kyle Polich : Definitely it's
Kyle Polich : Rather ubiquitous I think everyone has personal experiences both good and bad of recommender systems that worked great or sort of failed in maybe an embarrassing way but those are kind of our personal anecdotes from someone in your position where you get to see maybe the bigger picture when you consider all users and all products or all possible recommendations What sort of challenges emerge as you pursue it from the scalability point of view
Boya : Yeah I think that when you talk about the scalability we should think about from two views One thing is from the user and item level which means that there's so many users and the items online right The items here are not limited to the the physical products we usually see on the e-commerce platform It can also include the the content product
Boya : You will browse uh many websites and sometimes that you search for a target website and browse that but sometimes you just randomly click something given your your your exposure at that moment
Boya : So that is the one aspect It's just from the user and item level But if we dive into the user and item themselves so that refers to the second level is that for one user he or she can be described by so many texts because we would have a lot of footprints online so far because of the frequent visits uh on the website and even some offline activities can be recorded
Boya : And in some way that they can be integrated with our online behaviors and also we have our basic demographics like where are you from like what gender and also what would be your uh education background and and so much information about one person or maybe a group of persons available right So for each user the scalability would about the dimensionality of the features used to describe this person
Boya : So the same case is also applicable to items If we think about a physical product that we can use price we can use uh a size we can use the ingredients to describe a food or color or producer location so many information not limited to a physical product If you look at the the content products such as the videos or audios that themselves are unstructured data
Boya : So when we talk about unstructured data that dimensionality is always not that small right And if you look at the video it will be about some visual elements and also would be about some like color contrast or some content design stuff even when we talk about the broadcast the audio product that we would have the the tune the emotion or or the pitch right There's so many dimensions to describe such a
Boya : Uh a content product So that is the second level of scalability It means that the demographic space or the attribute space of an item or the person is usually high dimensional So when we propose the algoririthm to match the user and content we need to consider both levels of the scalability from the user and item level and also from their attribute level
Kyle Polich : What are the traditional approaches people use when they first went to tackle recommender systems like what are the classic algorithms
Boya : It's pretty hard to define a classical algorithm because you know for the high technology like that especially the digital technology
Boya : The technic itself is always like moving forward And so far it has been 2025 and we know that ChaGBT this large language model has been so popular So if you ask about the classic method like there will be a lot of options and maybe some advanced one but if you ask me this question three or four years ago my answer would be different So I would not name a typical one but I can tell you some
Boya : Intrinsic logic of dealing with the high dimensional data that is to do the data uh reduction So it means that in general no matter whether it's deep learning or maybe the collaborative filtering as uh the one discussed in my paper the core idea is the same is that we're going to figure out a latent space Such latent space is low dimensional one
Boya : But the low dimensional one can still summarize the core information that distinguished items or users in their original high dimensional space of features and uh demographics and attributes It it seems that we find a low dimensional space to summarize
Boya : the important information in the original high dimensional space So that is the logic So if we are talking about the deep learning that the terminology we usually use uh for the low dimensional vector in the low dimensional space would be embeddings
Boya : Even for the unstructured data people can transfer the unstructured one to the embeddings that is a low dimensional vector of numbers right The same case here if we talk about the collaborative filter or the matrix factorization so the terminology is usually the factorization the factors the low dimensional factors that can summarize the information in a high dimensional original vectors
Kyle Polich : I think that's an excellent perspective that whether it's collaborative filtering or deep learning or some technique that's gonna come out next year we need a process to put data into a fixed vector into that latent space and maybe your mileage will vary on which is the right approach for you Once I have that latent space where I've kind of have a I guess a feature vector on every item can that help me with the cold start problem
Boya : Yeah the coar problem is uh pretty salient especially when the for like when a website doesn't have any information about a new guest right Because as I mentioned that people can randomly browse many pages like online pages
Boya : Every day But if we think about the low dimensional that I mentioned when you have such a low dimensional space you you hope that the low dimensional space can include the information predictive of people's preference but how it could be predictive
Boya : So the generation or how we can figure out such a predictive low dimensional space depends on how much we have known about the focal users or the focal items But if I don't know about him at all maybe it's the person's first time to visit How can I generate such an informative uh low dimensional space That is the challenge of code star Coldar means that I don't have enough information that can help me to generate such an informative low dimensional latent space
Kyle Polich : Yeah it makes sense that OK if I've been on the site for a long time and I've given you a lot of feedback I should have an expectation that you've maybe learned me a bit better Could you describe that process how do we go from zero to some model of the user
Boya : So that is related to how we design our methods So the point is that people could have a potential for a user to have a group of counterparts who share similar preferences Right If we know something about the group of counterparts or people who are similar to the focal new users maybe we can use other people's behaviors to inform
Boya : The potential preference of this new user but this initialization would not be accurate but at least it could be informative
Boya : So that is the logic is that uh we don't use the focal new users' behavior to generate such a latent space because actually we don't have this information But it's hard for us to have no information at all because at least you can have the the person's location right That is usually together with the cookie or there's so many ways for a website to to have some basic information that is the basic demographics
Boya : And then we can use this basic information and link it with another group of users who would be similar to the focal new user and then we can use other people's preference to initialize uh the recommendation to the new user It means that we estimate a prior a prior space of the low dimensional representation for the new user and to handle or the mitigate the coldar problem
Kyle Polich : Makes sense yeah I like you described it as a prior as well cause now I'm picturing some Bayesian process or something like that where as information trickles in you can update it Do you have a sense of how long something like that takes I can't imagine if I rate one movie suddenly you know everything about my movie taste What does it take to get bootstrapped
Boya : You're asking about how long do I need to really understand a new person's preference maybe around several rounds of interaction with him and the longer would be uh let's say 5 rounds and the shorter one would be just the 1 round as you mentioned where you only interact with one movie and I can totally understand you as an algorithm
Boya : So I think that uh how long it will take depends on how different the focal user compares to the existing user that I can match with the new users or the people that I have known
Boya : From this view that you can categorize the new user into two groups One is that the user is just a normal one Means that it's a is a member of a majority Let's say Mhm Uh the person shares the the common interest with maybe 90% of the population here
Boya : But the other case is that the person is pretty niche She has a lot of niche preference that we don't have good representation of her type in our training data or even in in our real world For the first type the majority one in general that the learning of the person's preference could be pretty fast because our priors can be very accurate more accurate than the priors for a niche user
Boya : But for a niche user I think that how long do we need to spend learning the person's well depends on the learning process that we design in the recommendation system That is the reason why in our method that we also incorporate a banded learning That is the effective way of learning by doing
Kyle Polich : Yeah I'm thinking of cases like YouTube where I don't know the statistic but they're getting some absurd amount of videos uploaded per day and some of those new videos are going to be very popular but maybe a lot of them are are almost garbage they probably shouldn't have been uploaded you know the average video probably isn't a good video How do you balance the process of exploring trying to find a good video versus saying OK we know enough now this is either a good one or a bad one to be promoting
Boya : That's also related to the prior or the or the starting point of your exploration right Suppose that you start from a random place it means that you don't have much prior knowledge about whether the new user or would like or what kind of video but you just randomly select
Boya : Suppose that you have 100 videos so far and the very naive way is just to randomly pick up like several or maybe one and show it to the person and see his feedback right And then you use his feedback to update your your your algorithm for the for the future rounds right But because it's a random pick
Boya : It is very likely that the person would not like that and you usually have a pretty long uh stage of learning But think about the case that you are the user If you cannot find the movie you like on the platform for a while and what you're going to do you're going to leave this platform and switch to another platform So the consumer churning would happen
Boya : So it's always important for the firm to learn about the user's preference in a short time and to increase their retention rate to to keep the the customers in their platform So the solution is like the high-level intuition that we need to have a good priors And we know even if the prior is not accurate to capture the real preference but it can help us to narrow down the target areas that we
Boya : We should search for more feedback and improve our algorithm If you start from something wrong and it will take longer for you to go to the destination or even maybe forever you cannot arrive in the destination right But if you start with something right or maybe that's the area you explore it's just around the true preference of the person and then you will definitely take some shorter time to get the point
Kyle Polich : Yeah I guess if you did a very naive implementation of multi-arm bandit on YouTube I would end up much of the time seeing a video in a language I don't understand uh which most of the time would be a skip So of course it would be good to have another solution like collaborative filtering to kind of pick up the balance there but how do you get these two married together
Boya : In our algorithmic design the first step is to recognize the low dimensional space by linking the new user with some existing group of uh consumers given the similarity between their demographics And we do the same thing to the item It means that for a new item that we would recognize a group of existing items that shared a similar attributes
Boya : Given this prior knowledge from both the user and item side and we can generate a relatively informative low dimensional space to represent user's preference on items But again because this is the initialized low dimensional space it is not accurate but it's uh it can be a good starting point for the exploration
Boya : The next step is to do the bandit learning Given the user preference represented by the low dimensional space we can choose the item that so far has a potentially good feedback from a user but it would have some uncertainties So in general that the objective function of the banded learning is not just maximizing the feedback it's
Boya : Actually maximizing the feedbacks uh plus a weighted uncertainty It means that we value the uncertainty So the goal is to reduce the uncertainty It means that during exploration we're going to recommend something that we're not pretty sure the new user would like but we know if we can collect the user's feedback to this item even if it's an active feedback it will highly help us to reduce our uncertainty
Boya : of understanding the user's preference It means that the information of uh getting the user's feedback to the recommended item is informative and then we'll recommend the item After recommending the the item and we observe the user's feedback and we include the user's feedback into the algorithm to do a Bayesian updating and then we would have the posterior low dimensional space
Boya : At the end that our like algorithm estimate of the user preference would converge to a a stable one and it's usually a good or the accurate estimate of the user preference
Kyle Polich : So I know Bandit is typically an online learning technique and uh of course you could take something like MovieLens and simulate something online I suppose but could you talk a little bit about the experimentation or your approach to either testing on a platform or simulating something like that
Boya : In our paper that uh as you mentioned that we do have some empirical testing on different data sets that the first is the standard and archive one but as you mentioned that standard learning is offline one if we use the offline data to evaluate that the bias is inevitable and we also have some synthetic data but it's still the offline one
Boya : But at the end we conduct our online experiments on an online grocery setup It means that we ask the participants the the real human being to join our experiment that they do the shopping as what they do on Amazon or Instacart
Boya : Given their feedback to the recommended products we list on the product page we update our algorithm and then we have multiple rounds of experiment It means that in the 2nd session the 2nd round and the same group of participants will be invited to do the 2nd or
Boya : Second the shopping again and the recommended items to each user would change The change depends on their feedback in the previous run So we keep the algorithm updating in a life mode So it's the same as how you interact with Amazon or Instacart
Kyle Polich : And can you share a few details on the results
Boya : We in the experiment that we benchmark our method with another 3 algorithms that each of them doesn't have one element compared to our method So remember that our method have the collaborative filtering that is the basic one that usually conducted for the user's feedback to items but in addition to this component we have the collaborative filtering on the demographics and the attribute this two-side matrix to inform
Boya : The low dimensional space when we don't have the user feedback to solve the code star and the third component is the banded right So we benchmark this method with another 31 and each one doesn't have any one of the components And overall our method uh outperforms these 3 alternatives on average that is the average result
Boya : Second we try to figure out the the source of the benefits for our method by comparing our method with each benchmark for the category of products such as protein Then people's preference highly rely on their demographics which means that the priors themselves can be very informative So the banded learning so far is not that important
Boya : In this case we can find that our method doesn't have a significant difference compared to the performance of the method that doesn't have the banded learning However in some other categories such as produce I think people's preference for sure are correlated with their demographics but not highly relied on that It means that the demographics can only provide some information but not all the information about their preference
Boya : So in this case banded learning matters a lot because that the banded learning can help us to collect more information which is built on the initial low-dimensional space So it can include a lot of valuable information that the demographics and the item attributes doesn't have at the very beginning So in this case we can find that our method can significantly improve the performance compared to the method without the banded learning
Boya : And overall we find no matter which scenario it is the data reduction is pretty important
Boya : That gets back to your question at the very beginning about the scalability So we know that learning from the original features which is a high dimensional one is slow That is the first thing Second that the information integration is not good So it they cannot provide uh the information uh predictive of the people's preference because you usually need the synergies across the multiple information
Boya : Is that not the additive form of uh uh each small piece So you need you need to have a model that can well integrate all of them That is the informative data reduction
Boya : That is the second takeaway The 3rd takeaway is that we compare the benefits from our method across two user groups One is the majority one and the uh the other is the the the minority one The minority user are those people who have the niche preference It means that they like something unusual like they deviate from the mainstream or the major population's common interest
Boya : For this group of people we find that our method because of the banded learning can do better than the method that doesn't have the banded learning The reason is that
Boya : The collaborative filtering that is a good dimension reduction method but because the the majority usually has a lot of data in the training set right So they are well represented So for them learning about their preference can be very fast and the priors for them can be very accurate even at the very beginning So they always benefit from such uh information
Boya : Integration but for the niche users that the platforms or the algorithms understanding about their preferences are so uncertain is because their data their preference cannot be well represented in the existing users So the learning helps a lot Because of the learning that the difference between the benefits for
Boya : Majority and minority can be mitigated It means that our algorithm looks more fair to both groups So that is the three main takeaways about our experimental results
Kyle Polich : Well I consider myself to be one of those niche interest people in that small group so I'm appreciative of any algorithm that can service me as uh I guess a minority in that case I'm wondering if we could put ourselves in the shoes of that new user then so they come to whatever platform that let's say is powered by your approach and uh they've got to give a little bit of feedback that will then place them in that latent space so you can kind of get a similarity if I understand correctly
Kyle Polich : But that's an unsupervised process so it's not like you can tell me exactly what it is without using the word eigenvector right it's sort of just a mathematical answer but I as the user I'm going to experience it and maybe I'll say oh it learned the genre of music I liked or it learned the time period I like or it learned something about me Do you have any sense of interpretability between the vector and uh the user experience
Boya : The interpretability of the user experience and the latent space or how the algorithm matched the observed the demographics or item attributes into the low space is very context dependent
Boya : So far we we use our method in the grocery shopping and uh we we can find that some low dimensional vector because it's just a vector it's it's actually a combination of multiple observed demographics and items could be different for each product categories For example that first
Boya : Some vector in a low dimensional space we can name it as the nutrition oriented one because that low dimensional space is highly relevant or correlate with people's nutrition-related diet or their age But then if we apply the same algorithm into another product category
Boya : Maybe the beverage and then you would have different interpretability of the low dimensional space and their linked with the observed the user demographics and item attributes So there is uh no specific answer about the interpretability because it's completely context dependent
Kyle Polich : I've touched a little bit on the topic of fairness in some other interviews I guess could you give us the high level of how it would plug in with your approach
Boya : In my research so far that I think my understanding about the algorithmic bias is more about whether the same algorithm would be different for different groups of users If I'm a niche user why should I always spend a longer time interacting with a website to get what I like compared to people belonging to the majority right It's kind of some bias and that is what we mean by unfairness
Boya : So I think about the problem from first the the the biasedness in the training data that most algorithm is using even if we have the unsupervised learning but uh in practice the good performance is usually come from a semi-supervised or some supervised learning So we we need some guidance from the existing data right But what if the existing data is already biased So different group of people can be represented differently in the in the existing data
Boya : And also the accuracy of this representation can be different My my research so far is trying to say that how the common component in the algorithm design leverages the existing data and as a result how the intrinsic bias in the existing data can be transferred to the algorithmic outcomes and people can feel that in their user experience Whether the solution should be improved
Boya : Moving the data or whether the solution would be more about improving the algorithm design to have a better use of the existing data Because the people like the the proportion of different group of people is definitely not even That's that's for sure right That's the reason why we always have majority and we always have minority But given the the data is just like that that is the fact and how the algorithm design can correct it or can avoid some bad outcomes that violates the fairness
Kyle Polich : Well I know this paper's a good milestone in and of itself and uh it's just a slice of your overall research but are there any next steps along these lines
Boya : Yeah along these lines as I mentioned that this is an empirical paper so far and we're trying to have some like more solid exploration from a theoretical view and to show is there any like close form solution of the
Boya : Effects from different uh algorithm design on the fairness outcomes and what would be the right or not right what would be appropriate matrix of evaluating the fairness or the algorithmic bias when we include multiple uh groups of users or
Boya : Multiple group of items uh with different uh market shares or the proportion in this market That is more like a theoretical uh exploration along this line and another line would be about the effect of the recommendation system on social media The reason is that the recommendation doesn't only affect the user experience I mean as from a demand side it also affects the supply side
Boya : So suppose that you are a creator or in some cases you are an influencer How to choose the topic of your video Mhm It depends on whether you want to get more views It means exposure or whether you want to get more positive feedback Maybe that we can use like to quantify the positive feedback
Boya : For sure that the number of views or the number of likes that depend on your video quality but it's it's not limited to that it depends on who is going to receive your content
Boya : If your content is recommended to someone who already have some positive bias about that which means that they are more likely to like it and then it's also more likely for them to give you a positive feedback But what if your video is recommended to somebody who is more irrelevant even if your video quality is high enough and uh you still have like less opportunity to be liked The problem is
Boya : You are not the only creator on this market There are so many other creators and uh from the demand side of the audience attention is limited Everybody only has 24 hours
Boya : So how to compete for the attention it depends on what other content are recommended together with your content to a focal user So that is the next step to understand how this uh traffic allocation the recommendation system manipulate or affect the content supplier's decision about what topic to produce
Boya : Because you can choose the popular topics to to produce it's more likely for you to get exposure but also it's always easy to get exposure for others who produce the popular topic But the problem is that the market for this topic could be more competitive
Boya : And how likely can this small creator get more attention But if you produce something so niche well the recommendation system cannot find good or the relevant audience to watch the video and still you have some risk of having a very small or positive feedback
Boya : But then how to balance that At the end you may find there could be some equilibrium decision on the topic choice And then maybe on this platform a lot of content only focused on the popular topics It means that the content centralization is is so severe but maybe if you use another
Boya : Recommendation algorithm or the logic and then the the platform the distribution of the content on the platform can be very diverse So polarized content versus diverse content can also be an outcome of the recommendation system So that would be the potential effect of the recommendation system on the supply side So I would focus more on the effect on the supply side in my future research
Kyle Polich : Yeah I think that's rather novel I'm not aware of a lot of good work in the supply side space I suspect a lot of creators are desperate to get algorithmic feedback over just you know uh things that might even be speculation that they're getting now on how to make the right video or or whatever they're producing
Boya : Yeah because I chat with some influencer friends they told me that even for the same piece of video it's really hard to get popular on every platform because the like different platforms have the different logics of uh allocating the traffic through their recommendations
Boya : System even if the video quality is the same even if that the audience base is also similar across some platforms but still from the creator's view that the performance of their creation can still be so different it all depends on how the content can be recommended
Kyle Polich : Boy thank you so much for taking the time to come on and share your work
Boya : Yeah thank you so much for the interview and still that's a good opportunity for my work to be known to the public Yeah thank you
Kyle Polich : Definitely