Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
SlideShare a Scribd company logo

The Data Driven Company

•
•
Bart De Waele
Bart De Waele

This document discusses how Shein uses data and algorithms to disrupt the fast fashion industry. It summarizes that Shein uses an engagement algorithm in its app to gather data on customer preferences and only produces items that generate high engagement. This allows Shein to produce smaller batches tailored to customer needs with less capital and faster production. The document advocates that companies flip their production model to first generate customer data and then produce items, with the user experience designed to optimize data generation and a self-learning algorithm that improves the experience over time.

1 of 25
Download to read offline
The Data Driven Company
2 Every board room ever… “What are we doing with DATA?”
3 Data is the new oil, right?
4 Data is like sand… All around us. Cheap and even without value of itself. Unless you use it in your production process.
5 And I’ll use this as an example:
6
7 Chances are, unless you have a teenage daughter, 
 you won’t know this company.
8 But fi rst, a little detour. (And yes, you also need a teenage daughter to know this app.)
9 TikTok 
 disrupted social media, by using a new kind of algorithm based on engagement data. Not determining reach and spread of content based on our social network (number of ‘friends’ we have) but on the engagement each piece of content generates.
10 100 high engagement low engagement 1.000 drop high engagement low engagement 10.000 drop high engagement low engagement 100.00 0 drop Better ‘ fi t’ for viewers > addictive app.
11 TikTok: algorithm. Engagement data in the core of the product. Thus delivering a better user experience. And using this feedback loop to create a better product.
12 Back to Shein. Shein is the TikTok of fast fashion. Using an engagement algorithm 
 to disrupt fashion retail.
13 100 high engagement low engagement 1.000 drop high engagement low engagement 10.000 drop high engagement low engagement 100.000 drop Showing products to a limited set of users, and using digital engagement as decision engine to produce or not.
14 Data as start of the product. Using engagement data in the app to predict the success of fashion items. Consequences: • better fi t to user need as engine learns • only producing what is virtually sold, creating less stock • predicting raw material need, needing less capital • faster & smaller batches, allowing vastly more SKU • serious price advantage, only improving with algorithm
15 design & produce items (‘guesstimating numbers’) ‘Normal’ fashion companies: distribute to warehouses & stores marketing & sales gauge digital intrest Shein: produce exact number of items ship directly to consumer end user end user
16 Flipping the supply chain. 
 Putting user experience & data fi rst. Creating a disruptive advantage.
17 Caveat! • I am not endorsing fast fashion! • Shein is also exploiting loopholes in trade agreements & taxations. • Always be wary of Chinese companies & privacy. • Not easy to integrate your total supply chain. • D2C not always an option.
The Big Flip What can we learn from Shein? 1. Flip your production chain 2. UX as data generator 3. Self-learning engine
19 Flip your production chain. First generate the data of what your customers will buy, and only then produce the items. 1
20 UX as data generator. • UX of the Shein app is mostly aimed at generating engagement data and insight in your taste. • Conversion to sale comes second. • As a new user, your engagement data is more valuable 
 in powering the fl ywheel then your initial sale. 2
21 The algorithm catches user data, which makes the user experience better, which attracts more users, which generate more data. 3 Create a self-learning loop.
22 Data in your core. Data is not something that is generated as a by-product of your core business; 
 and afterwards tortured to generate some magical insight. Data should be embedded in the core and at the start of your production line.
23 Questions to ask yourself. • Can we generate user data? • Can we put that data at the start of our production process? • Can we build a self learning feedback loop to enhance the user experience? • Can we engineer the user experience to generate more and better data? Build a data driven company.
24 Questions to ask Duke & Grace. How can Duke & Grace help you engineer your user experience? bart.dewaele@dukeandgrace.com
25 Questions? Shoot.

More Related Content

The Data Driven Company

  • 1. The Data Driven Company
  • 2. 2 Every board room ever… “What are we doing with DATA?”
  • 3. 3 Data is the new oil, right?
  • 4. 4 Data is like sand… All around us. Cheap and even without value of itself. Unless you use it in your production process.
  • 5. 5 And I’ll use this as an example:
  • 6. 6
  • 7. 7 Chances are, unless you have a teenage daughter, 
 you won’t know this company.
  • 8. 8 But fi rst, a little detour. (And yes, you also need a teenage daughter to know this app.)
  • 9. 9 TikTok 
 disrupted social media, by using a new kind of algorithm based on engagement data. Not determining reach and spread of content based on our social network (number of ‘friends’ we have) but on the engagement each piece of content generates.
  • 10. 10 100 high engagement low engagement 1.000 drop high engagement low engagement 10.000 drop high engagement low engagement 100.00 0 drop Better ‘ fi t’ for viewers > addictive app.
  • 11. 11 TikTok: algorithm. Engagement data in the core of the product. Thus delivering a better user experience. And using this feedback loop to create a better product.
  • 12. 12 Back to Shein. Shein is the TikTok of fast fashion. Using an engagement algorithm 
 to disrupt fashion retail.
  • 13. 13 100 high engagement low engagement 1.000 drop high engagement low engagement 10.000 drop high engagement low engagement 100.000 drop Showing products to a limited set of users, and using digital engagement as decision engine to produce or not.
  • 14. 14 Data as start of the product. Using engagement data in the app to predict the success of fashion items. Consequences: • better fi t to user need as engine learns • only producing what is virtually sold, creating less stock • predicting raw material need, needing less capital • faster & smaller batches, allowing vastly more SKU • serious price advantage, only improving with algorithm
  • 15. 15 design & produce items (‘guesstimating numbers’) ‘Normal’ fashion companies: distribute to warehouses & stores marketing & sales gauge digital intrest Shein: produce exact number of items ship directly to consumer end user end user
  • 16. 16 Flipping the supply chain. 
 Putting user experience & data fi rst. Creating a disruptive advantage.
  • 17. 17 Caveat! • I am not endorsing fast fashion! • Shein is also exploiting loopholes in trade agreements & taxations. • Always be wary of Chinese companies & privacy. • Not easy to integrate your total supply chain. • D2C not always an option.
  • 18. The Big Flip What can we learn from Shein? 1. Flip your production chain 2. UX as data generator 3. Self-learning engine
  • 19. 19 Flip your production chain. First generate the data of what your customers will buy, and only then produce the items. 1
  • 20. 20 UX as data generator. • UX of the Shein app is mostly aimed at generating engagement data and insight in your taste. • Conversion to sale comes second. • As a new user, your engagement data is more valuable 
 in powering the fl ywheel then your initial sale. 2
  • 21. 21 The algorithm catches user data, which makes the user experience better, which attracts more users, which generate more data. 3 Create a self-learning loop.
  • 22. 22 Data in your core. Data is not something that is generated as a by-product of your core business; 
 and afterwards tortured to generate some magical insight. Data should be embedded in the core and at the start of your production line.
  • 23. 23 Questions to ask yourself. • Can we generate user data? • Can we put that data at the start of our production process? • Can we build a self learning feedback loop to enhance the user experience? • Can we engineer the user experience to generate more and better data? Build a data driven company.
  • 24. 24 Questions to ask Duke & Grace. How can Duke & Grace help you engineer your user experience? bart.dewaele@dukeandgrace.com