You work for a consultancy firm that recently developed a contract with Stitch Fix to advise it on its strategy development in innovation and breakthrough technology.

1. Discuss the key factors and strategic moves that enabled Stitch Fix success and competitive advantage in the previous years. Do you think the same factors and strategy orientations will be enough to guarantee the company success in 2023? What are the key challenges of the company success?
2. Discuss the key factors shaping scaling of Stitch Fix in the previous years. Can the company continue relying on these in 2023 or would you suggest considering newfactors?
3. What ethical and business challenges might Stitch Fix face when developing its AItechnology?

\fWhether you are a Shopaholic, a need-based shopper or totally disinterested by retail shopping, you will need to shop for clothing. Everyone must. In 2018, the US. apparel market was valued at approximately US $340 billion, US. dollars,1 which makes it the largest in the world. Stores have always been the primary source of sales in the U.S. Apparel industry, but while most shopping is still done in brick-and-mortar retailers. online sales had grown faster than brick-and-mortar sales and \"Apparel. Footwear and Accessories" is now the largest U.S. e-commerce category.2 When you experience brick-and-mortar shopping. you can see and touch products as well as take items home immediately. But in-store shopping is time-consuming and inconvenient: go there, park your car, enter crowded shops, try on clothes in tiny rooms with questionable privacy, get limited choice that is typically far more expensive than online and have limited opening hours and long check-out lines. These pain points have propelled the Apparel and Accessories e-commerce revenues to US$103 billion in 2018 rising to an expected $145 billion by 2023.3 Low price and convenience' rank highest among the reasons why customers choose to shop online versus in stores and most online stores focus on low prices, fast delivery and easy returns. However. online shopping for clothes has its downsides as well. Because of the tremendous choice, online shopping has become time-consuming and confusing. Customers have countless websites and an endless number of pages. This overwhelming selection of apparel, shoes and accessories, coupled with a lack of professional advice and the impossibility to assess quality and how an item actually ts before buying can make the online shopping experience difficult and frustrating, especially when customers do not know which style suits them. What used to be extremely convenient in the rst place has become tricky. Stich Fix: Reinventing the E-commerce Retail Experience "Spending a day at the mall, or devoting hours of time to sifting through millions of products online is time consuming, overwhelming and neither effective nor enjoyable. I knew there had to be another way\" Katrina Lake, CEO of Stich Fix Stitch Fix, a San Francisco-based ecommerce company, is a personal styling service that chooses clothing for shoppers using a combination of artificial intelligence (Al) and professional stylists and sends a personalized box of apparel based on the customer's sizes and preferences. In 2017, founder Katrina Lake, named by Forbes as one of America's wealthiest selfmade 1 Apparel , Unlted States | statista Market Forecast, (n.d.). Retrieved from https llwww.statista.com/outlooleDOODOOO/t09IaggareVunited-states 2 Garcia. K. (2019, February 11). The State of us Apparel Shopping in Five Charts. Retrieveo Febmary 25. 2019. from httos l/www.emarkelercom/conlentllhe-slate-ol-us-aoparel4hoppinq-ln-vecharts 3 Us. online apparel and taehten market size 2022. (n,d.). Retrieved from htips //www soattsta eomtststtsites/27ssso/us-aeperei-ertdaceessortes-retatle-eommereerevenuy 4 Total Retail 2015: Retailers and the Age of Disruption, (2015, February). Retrieved February 2019, irorrt ti . o o i tio to - t -2 Copyright INSEAD 1 This oocuntant ts authorized tor use only tn Aljona Zolina's Strategy and organization or Innovation A2 Nov 22 at lESEG School ot Management lrom Nov 2022 to May 202: r i'lllt women,5 was one of the few women to lead an IPO and at 34, Lake was the youngest. That year less than 8% of IPOs were led by women; since 2000 the percentage has never exceeded 13%.5 The Stitch Fix adventure started in 2010, as part of a class project, when Katrina Lake tested the concept, formerly known as Rack Habit, in her apartment, while finishing an MBA at Harvard Business School. Lake had a lifelong love of clothing; retail fascinated her. But she couldn't help noticing that retail, an age-old industry. never changed, She just kept thinking about how technology could be leveraged to change the value you are able to provide to people and. in doing so, create a fundamentally different and superior experience to buyers. Lake had watched industry leader Blockbuster quickly succumb to Netix. a company that leveraged technology to create a leap in buyer value which set it apart and was irresistible to people, In 2009. Netix pivoted from mailing disks to streaming videos, creating an entirely new experience. Customers could not only stream a rich array of movies but also amazing original series like Crown, House of Cards, and Orange is the New Black, anytime, anywhere, anyway without leaving the comfort of their home, a new value proposition. In contrast, retail may as well have been selling 1970s leisure suits. There were so many inefficiencies. Onltne clothing stores just basically digitallzed the physical stores, creating more selection but removing the ability to try and buy immediately. This was not a new value proposition. How would technology possibly change that, Lake wondered? Lake thought about her sister, a personal shopper in New York City who helped women find the right clothes to match their attitude. body type and sense of style, and wanted to extend that privileged experience to women everywhere. But she knew that only the wealthy and the in-the- know would go for a personal shopper, such as Michelle Obama who is famous for always looking great thanks to the personal shopper she uses. Lake also recognized that people don't have time anymore and widely enjoy the convenience of online shopping. hence explaining its enormous growth. She wondered how to combine the best of online retail stores the large selection and the ability to shop from anywhere, anytime with the ultimate tailored experience of personal shoppers. Doing that. she reasoned. would offer a leap in value and open up a new market. But Lake's reasoning didn't stop there. At the same time, in her personal life, Lake had joined a community-supported weekly fruit and vegetable subscription where customers received seasonal produces directly from local farmers. Because these boxes depend on what the farmer harvested, customers don't know in advance what it would contain, They only know that it will be fresh and in season, so likely delicious. Lake loved the surprise factor and that it would push her to try new fruits and vegetables she had never tried before. Then the idea came up: why not do that with clothing too? Making It Happen To make this work, she knew technology would be a key enabler. Was there a way to combine the human element of a personal shopper with Al doing the early matching so a business could be built that not only would customers love but would importantly also be economically scalable? 5 O'Connor, C. (2017. November 30). Stitch Fix Stack Surge Puts CEO Katrina Lake Among America's Richest Self- Made Womsnt Retrieved from httbs'l/WWW.turbeSLomlsites/clareoconnor/Zm7/11/2Blslllvil-x-stuck-Surde-DutS-oeo- katrtna-lake-amo"(Iramerlcas-rlchest-Selfrmsde-womenljlbeBeBSOVS 6 Rey, J D, (2017, November 17). Women in te it t day's St't h FIX IPO as a milestone, Retrieved from h ' ox 0 1 66 k - - ou Copyright INSEAD 2 This document is authorized tor use only ll'l Aliens Zorina's Strategy and organization at Innovatlort AZ Nov 22 at IESEG School ol Management lmm Nov 2022 to May 2m. Language Processing (NLP). a sub-field of Al that enables computers to understand human languages. Orders are then routed to the styling algorithm. First. the algorithm removes from the inventory list all clothing either already sent to the client (purchased and returned items) or marked 'undesirable' by the client. Then the machines identify items likely to delight customers. Each item is subject to 100-150 data points (colour. material, length, etc). This data, coupled with algorithms and customer feedback, allows Stitch Fix to calculate a probability score of purchase for each item. Moreover, Stitch Fix has photographic data on all its inventory that trained Articial Neural Networks\" turn into vector descriptions. As 46% of Stitch Fix customers have a dedicated Pinterest board with clothing inspirations,\" these trained Artificial Neural Networks also turn Pinterest inspirations into vector descriptions and then calculate the similarities to optimize the selection. Once done. the shipment request is routed to a stylist. Even the choice of the stylist depends on an algorithm. To be associated with a particular customer, a score is calculated based on the availability of the stylist. but also on a previous relationship between a stylist and the customer (based on previous orders) and the affinities between the customer's style preferences and those of the stylist. Once determined, the Stitch Fix stylist selects five items (clothes, shoes and accessories) from the list produced by the algorithm by reviewing customer preferences in the style profile. Finally. a warehouse worker scans and collects each item successively. collecting one item after another, just like a personal shopper would do. Because Stitch Fix is constantly growing, its engineers developed warehouse maps and an algorithm to optimize the routes workers take; this reduces time and error.\"2 On the scheduled date, customers receive their Fix without knowing what is inside, which adds a little \"surprise" factor. In addition to the five tailor-picked items. the customer finds a Style Card for each item in the box that gives styling advice (what to wear with each item. how to mix it with the proper shoes...etc.) and a little personalized note from the stylist explaining why s/he picked certain items based on the customer's prole recommendations. The customer decides what to keep knowing that there is a 25% discount where everything is kept and what to return. They are also asked for online feedback, identifying what s/he liked or disliked, to help plan future Fixes. Customers can invite friends by sharing a referral link; they receive a $25 credit for every friend who orders their rst Fix. In February 2018. Stitch Fix went a step further and launched a new service. Extras. that allows shoppers to add extra products to their Fix such as underwear and socks from either top brands or Stitch Fix private label Everyday. It's a great way to up-sell essential products and discourages 4o Articial Neural Networks are one at the main tools used in machine learning As the "neural' part or their name suggests. they are brainelnspired systems which are intended to replicate the way that we humans leam. Neural networks consist or input and output layers. as well as (in most cases) a hidden layer consisting of units that transiorm the input into something that the output layer can use. They are excellent tools for nding patterns which are far too complex or numerous iora human programmer to extract and teach the machine to recognize. What is an aninoial neural network? Here's everything you need to know. (2019. January 06). Retrieved from htlos:/lwww dioitallrends com/cool-tech/what-is-an~arlllicial-neuralrnetworkl 4t Faguer. L. (2017 May 22). Comment Stitch Fix exploite le pouvoirdes algorithmes ? Retrleved April 18. 2019. irom httpllcustomer slght onsulttngcoml22691l 42 Clark. J. (2017. Augus 1). The curious connection between warehouse maps. movie recommendations. and structural biology. Retrieved April 18. 2019. from ht Qs:I/multilhreaded,stitcnxccm/blmui7/08/31Iwarehouse-Iayouts/ Copyright INSEAD 7 This document is authorized lor use only in Alton Zorlna's Strategy and Organization or innovation AZ Nov 22 al IESEG School oi Management lrom Nov 2o22 to May 202:. in customers from visiting other stores to buy these undergarmenm, where they could be tempted to buy other clothes. These little \"Extras" increase the activity and the brand loyalty of Stitch Fix. Over time. Stitch Fix learns more about client preferences via detailed data on both the merchandise they keep and return. The more customers use Stich Fix the smarter it gets. Over 85% of the shipments result in direct client feedback.'3 This feedback helps Stich Fix not only improve personalization for customers but, also. with wholesalers and designers. Stitch Fix shares its ndings with vendors to better determine. in real time. what items will sell well, and which are less likely to be successful. The items available are all based on recommendations that emerge from customer surveys, Pinterest boards and personal notes to stylists. Information gathered is then used by algorithms, written by data scientists, to select the best items for the customers. Scaling Fast But Not Too Fast Launched as a women's-only service, Stich Fix now sen/es men, kids, petite, maternity, and plus size, but the largest portion of the company's revenue is still from women's apparel. By gradually expanding into newer categories, Stitch Fix is targeting the entire family through every step of their life (work clothes when starting a new job), men's clothing, maternity clothing. kids clothing. etc.). In 2013, 39% of Stitch Fix customers bought most of their clothing from Stitch Fix, up from 30% in 2015. Stitch Fix scales up fast. but Lake knows the key to sustainable growth sometimes requires slowing down. She wants to grow but refuses to expand at the cost of a poor customer experience. Lake paces Stitch Fix growth. Handling returns, making sure that every item is in perfect condition. and developing new product lines (men. kids, plus size...etc.) requires a lot of work. To thrive and grow, Stitch Fix does not hesitate to pull the brake. typically via the use of waiting lists. For instance, the algorithm once made a bad prediction about a forthcoming trend. What did Stitch Fix do? Instead of sending out suboptimal clothing, then suffering a high return rate and dissatisfied customers, Stitch Fix donated unfashionable clothing. explained the problem. and asked their customers to wait. What could have been a disaster ended up boosting customers' trust. As the business grew. they put potential customers on waiting lists rather than sign up people they could not adequately service. Stitch Fix achieved profitability in 2014, three years after its launch. Revenue grew from $73 million in 2014 to $1.2 billion for 2018.'5 '6 As of March 2019. it had 3 million active customers. yeareover- year growth of 20%.\"7 Stitch Fix works with more than 1000 fashion brands\" and also has its own 43 United States Securliies and Exchange Commission Washington. DC, 20549 Form S-1 Registration Statement. (2017. October 19). Retrieved April 18. 2019. from httgs:l/wwwsec.gov/Archivesledgar/oatal'l576942/000119312517313629/d400510ds1.htm 44 Faguer. L. (2017. May 22). Comment Stltcn Fix explolte le pouvotr des algonthmes 7 Retrieved April 18. 2019. from http://customerainsightrconsultingom/ZZBM/ 45 Understanding Stitch Fix' Finding the Perfect Flt. (2017. November 13). Retrieved February 24. 2019. from nttgs./lwww,gcodwatercag.com/thesis/underslanding-stitch-x 46 Pearce. D, (2015. October 01). Slilch Fix Announces Fourth Quarter and Full Fiscal Year 2018 Financial Results Retrieved February 20. 2019. from httns./llnvestors.siitchx cornews-reIesseslnewsrelease-details/siltCh-x announces-fourtn-guarler-and-iulI-ilscal-year-Zm8 47 Stitch Fix. (2019. March). Retrieved April 2019. from nttsz/investorS,stltcniix,com/statlcsflles/2b3986944553r45867 b763-e942617e4dbf 48 Quarterly Results. (n.d.). Retrieved from https:I/investors.stitchx.com/quarterlv-results Copyright INSEAD 8 This document is authorized tor use only in Aljona Zorina's Strategy and Organization ct innovation AZ Nov 22 at IESEG school oi Management irom Nov 2022 to May 2023, brands, which represented 20 percent of revenue in 2017. Gross margins increased from 35% in 2014 to 45.1% in 2017,49 5\" thanks to the launch of higher margin private label brands designed by and for Stitch Fix exclusively, based on customer feedback. In 2019, Lake was named among Fast Company/s Most Creative People in Business.51 Prots and sales are important but keeping customers engaged is vital for any online seller, Style Pass and Extras increase the retention and usage rate. To gather more customer data, which improves the algorithms ability to understand a customer's personal style, Chris Moody designed a feature in 2017 called Sty/e Shuffle, available on the Stitch Fix app since March 2018. This addictive like/dislike quiz allows customers to rate a set of clothing images each day while encouraging them to visit the app more often. With more than 75% of customers using the feature, Style Shuffle boosts Fix requests and keeps customers engaged. Another way that Stitch Fix found to keep its customers engaged is social media. Stitch Fix understood the value of its customers' social media accounts, especially Pinterest boards which focus on visual inspiration, particularly clothing. With Pinterest. Stitch Fix created an ecosystem of customers. A search for \"Stitch Fix" on Pinterest brings up dozens of pages of boards with hundreds of pins in each board. There are even tutorials on YouTube about how to create the perfect Pinterest board for your Stitch Fix stylist and over 12.000 videos of people unboxing their Fix, some of them with over 130,000 views. The algorithm allows Stitch Fix to learn about customer preferences and body types, and Stitch Fix gets to know who its customers really are with their Pinterest boards. Stitch Fix's own Pinterest account has over one million followers. They have so many followers because, unlike other retailers, Stitch Fix is not selling what they post on Pinterest; these are style inspirations, not advertisements. Besides word-of-mouth of women loving the service and telling their friends, and bloggers talking about it, Stitch Fix recently invested heavily in marketing efforts, using TV campaigns to attract new customers. In 2018, Stitch Fix spent $102.1 million on advertising, a 45% increase over 2017 and quadruple their 2016 advertising spend. In February 2019, Stitch Fix showed their new campaign to the US press, generating 60 million reactions on Twitter.\" Stitch Fix is endlessly growing and expanded into the UK in May 2019.53 That said, in July 2018, Stitch Fix saw the rst attempt by Amazon to jump into the blue ocean Stitch le had created: Prime Wardrobe. This "try before you buy" shopping service was free and available from the Amazon website to all Prime Members. Customers could select as many as eight items that they want to try. Then they have seven days to decide what they want to return or keep before their credit card is charged. Returns are easy with the included prepaid return label and a resealable packaging. Because customers have to select their items themselves. apart from the fact that they have seven days to decide what to keep before their credit card is charged, there was not much 49 Understanding Stitch Fix: Finding the Perfect Fit. (2017, November 13). Retrieved Febmary 24, 2019, from httgs I/www.goodwatercagcom/thesis/understanding-stilch-t'ix so tittpsl/wwwbloombero,com/opinion/articles/ZO19-01-03/stitch-fixs-successstorv-Is-underaDDreciated 51 ti o olggoti-mo-oio 201 52 Pasquarelli. A. (2019. February 25), AdAge. Retrieved February 25, 2019, from http:/Iwww.adageindia.inlmarketinolcmo-strategy/sittch-t'ixdebuts-newbrand-campaion-on-oscars- night/anicieshow/sa144799,cms 53 Jacobs, E (2019, May 10) Would you let inis woman choose your clothes? Retrieved from ti I n 7 >717 -11 - - 1 Copyright INSEAD 9 This document is authorized ior use only in Aljona Zorina's Strategy and Organization oi innovation AZ Nov 22 at IESEG sehooi oi Management ironi Niw 2022 to May 2023, difference with a usual Amazon shopping experience. That is probably why the service never took off: According to a 2018 Gartner survey among 1,500 Prime Members, only 3% used Prime Wardrobe.54 But in July 2019, Amazon seriously targeted Stitch Fix with their new service \"Personal Shopper," by Prime Wardrobe. With Personal Shopper, customers receive monthly curated clothing boxes of eight items, similar to Stitch Fix. Prime Members begin by lling in a survey detailing their measurements, style, and budget. This survey is then analysed by a mix of stylists and Al to select items that their customers will like. Customers can preview the items before shipping and indicate which they want to receive and which they don't like and should be changed. Personal Shopper is available exclusively for Prime members and costs $4.99 per month. Besides the cheap price, Amazon offers a large selection of brands and tree & easy returns. Amazon already has over 100 million Amazon Prime Members in the US, and the female members are all potential customers. Despite the fact that Stitch Fix exceeded all stock market expectations, in July 2019 the stock price dropped on the news of Amazon.55 However, analysts still recommend investing in Stitch Fix. They believe that the retail sector is so massive that Amazon's new service will not steal customers away from Stitch Fix and that it can even make Stitch Fix's offer more appealing. In the meantime, on September 12, 2019, in an attempt to lengthen widen and deepen its blue ocean and build barriers to imitation, Stitch Fix acquired the technology and intellectual property from Finery, another female-led company. Finery extracts shopping receipts from email and creates a virtual wardrobe from it. Suddenly, customers can have a picture of everything they bought. Finery comes up with recommendations for mixing the pants you just bought with a top in your closet for a great Friday night outt. It also reminds you about the return deadline, while monitoring for potential price adjustments of items you saw online but did not yet purchase. While it remains to be seen how Stitch Fix will use this technology, we could imagine that it will allow them to offer more mixing and matching combos for women so they know how to use the clothes that they have, whether they purchased them on Stitch Fix, or elsewhere, Questions 1. Katrina Lake raised $42 million and struggled to nd investors to back Stitch Fix. Meanwhile, the male-run apparel retailing start-ups such as Fanatics, The Hut Group and Wish raised on average over $2 billion in funding. What role do you think gender played in this disparity/.7 How could Lake have more effectively tackled this issue? 2. Lake looked across three industries to create her new market space: retail, personal stylists, and subscription boxes. What did she take from those three industries and what did she leave behind? 3. From the buyer's perspective, what are the unique ways which Stitch Fix offered a leap in value beyond other retailers (brick-and-mortar and online)? What do you think will be the likely impact 54 Chen, Cathaleen, "Is the Stitch Fix Model the Future of Amazon Fashion?" The Buslness of Fashlon The Business of Fashion, July 31. 2019. ntiDs:Ilwww.businessoashion.comlarticlesews-ana| sis/amazon-DersonaI-shODDEr-SMIst x s F c L (2019, September 5) s h h I in | h m i 4 11- ' 55 "reasonably Beaten Down, Retrieved from ni Copyright INSEAD 10 This document is auihonzed ior use oniy in Aljona Zorina's suategy and Organization oi innovation AZ Nov 22 at IESEG sehooi oi Management ironi Nov 2022 to May 2023, of Amazon? What advice, if any, would you give to Lake concerning Amazon's \"personal shopper\" offering? 4. Will Stitch Fix's recent purchase of Finery IP help to act as a barrier to imitation and extend its blue ocean, especially against Amazon? With the understanding of what Finery does, how can you imagine Stitch Fix could use this to really add value for customers? 5. If you look at Stitch Fix's prot proposition, what do you see as key challenges to growing revenue and prots and protecting the blue ocean it created? Copyright INSEAD 11 This document is authorind tor use only in Alisha Zmlnl's Strategy and organization oi Innovation AZ Nov 22 at IESEG School or Management lroin Nov 2022 to Mary 202:, Appendix A The Algorithms Tour Unlike Facebook or Google that closely guard their algorithms and keep them secret, Stitch Fix created an interactive tour aiming at sharing how some of the diverse applications of articial intelligence at Stitch Fix work, On top oi the algorithms already described in the case study (warehouse assignment algorithm, intelligent logistics and the machines performing a variety of algorithms to produce rank-ordered lists of the inventory). Al is deeply engrained in everything that Stitch Fix does: in recommendation systems. in human- based computation55 resource management, inventory management, algorithmic fashion design and many other areas. Client' \"States" To anticipate the needs of their customers, Stitch Fix has to know the "slate" of each client at each point in time (new client? Did they use a referral link? Are they going through a new phase in life. etc.) , as they will likely have different shipment cadences, different desires for email contact, etc. Stitch Fix keeps track of every touchpoint with each clientevery item sent, feedback given, and every email using this data to understand clients' states and related needs. They can detect changes in state and consider possible triggers. Once the state is dened, they can develop state transition matrices and Markov chain57 models that allow them to study system-level effects, and most notably to anticipate future demand, as inventory is often bought months before it arrives at the warehouses. Inventory Management Stitch Fix must anticipate its clients' needs to have enough of the right resources in place at the right time. To do so, it uses system dynamics58 to purchase and maintain adequate inventory availability for stylists while minimizing the sum of ordering costs; meet future demand, allocate inventory appropriately to different warehouses, and occasionally donate old inventory to make room for new styles. New Style Development Volumetric challenges are not the only considerations with inventory replacement. Stitch Fix also want to purchase and develop new clothing in ways that continually improve its inventory, helping their stylists give greater delight to a broad client base. Designing new styles for their Stitch Fix exclusive brands provides a great opportunity to tailor new designs for particular client segments that tend to be underserved by other brands. Stitch Fix approaches this with inspiration from genetic algorithms59 that use recombination and mutation along with a fitness measurethe same mechanism used by mother nature in evolution by natural selection. The rst step is to think of each style as a set of attributes ("genes") Then consider the vast set 55 Human-based computation is a computer science technique in which a machine performs its iunction by outsourcing certain steps to humans, 57 A Markov chain is a stochastic model describing a sequence of posslble events in which the probability oi each event depends only on the state attained in the previous event, 56 System Dynamics is a tool to investigate and model complex dynamic problems in terms of stocks (the accumulation oi things), ows (the motion oi things) and feedback loops at any level oi aggregation. System Dynamics is now one oi the most commonly used lorrns oi computer simulation for dealing with many iaceted problems. The objective oi this method is to rind the conditions under which a system under study will evolve and in what direction. The tools used are models that represent symbolically the reality oi the system. it aims at considering the interrelationships between the components oi an organization or environment rather than looking at each component in isolation tttip://www.ioresight-platforrn.eu/community/torlearn/howtodo-ioreslghtlmethods/gaming-simulationand mgdeIsLsysem-dynam'cg 59 Genetic Algorithm (GA) represents a subset of Ignite Machine Learning APls. GA is a method of solving optimization problems by simulating the process of biological evolution. cAs are excellent for searching through large and complex data sets tor an optimal solution. Real world applications of GAs include automotive design, computer gaming, robotics. investments. tiafdshipment routing and more, https'I/agacheignttg reading toldogs/ggnettc-algonthms Copyright [NSEAD 12 This document is authoiized tor use only in Alton: zorina's Strategy Ind organization oi Innovation AZ Nov 22 at IESEG semi or Management trom Nov 2022 to May 202: of styles this way and consider the client feedback ("tness") available for each of them. Now consider creating new styles by recombining attributes from existing styles and mutating them slightly. Note that the number of possible combinations is very large (Flirt). In the next step, they deviate from a canonical genetic algorithm: instead of simply selecting based on fitness and then unleashing random recombinations and mutations as the next generation of styles, they are somewhat pickier about what makes it into their inventory. They first develop a model of how well a given set of attributes is likely to suit the target clients. They then use this model to highlight a variety of attribute-sets that they think have a high probability of being loved. Then they work with their human designers to vet and refine this collection, and ultimately to produce the next generation of styles. These new styles get produced and are made available to the styling algorithm, then on to delighted customers, and the cycle of evolution continues. Data Platform The Data Platform team provides the data and compute infrastructurealong with a collection of internal SaaS productsthat allow the data scientists to effectively and efciently carry out analysis, write their algorithms, and put them into production. The platform nicely encapsulates properties like data distribution, parallelization, auto-scaling, failover, etc. This allows the data scientists to focus mostly on the science aspect yet still enjoy the benefits of a scalable system. Engineers focusing on building platforms that enable data scientists to carry algorithm development all the way from concept to production. Source: A majority of the text in AppendixA is taken verbatim from: httpsjfalgorithms~tour.stitchx.com Copyright INSEAD 13 This document is authorized for use only in Aliona Zorina's Strategy and Organization of Innovation AZ Nov 22 at IESEG School ol Management from Nov 2022 to May 2023