No ratings yet.

Nowadays most businesses deal with some sort of data – their transactions, clients, product just to name a few types – and the larger the size of the data, the bigger the problems that arise. Large enterprises with long history and thousands of data points operate massive databases, build, developed and expanded throughout many years. Here’s where the real problems start – imagine your business, a Fortune 500-sized company, has gathered terabytes of data from multiple sources, which is stored in multiple databases, build by different teams and functioning in a very convoluted architecture. In the data there is a lot of hidden potential, but extracting it, combining and processing it to the point where it can be analysed and insights can be made, is a mammoth task.

Enter data warehousing companies like Snowflake which collect all the data, combine them into a single virtual place (“the warehouse”) and make it easy to run analysis on. It is compatible with pretty much all existing cloud databases (Azure, AWS or Google) and can shorten data extraction processes which used to take hours to a couple of minutes.

Snowflakes business strategy is unique in a sense that it sells a SaaS product that saves time and money (the cost of building an in house data warehouse is said to be over 350,000 USD per year [1]) to the largest companies in the world. The value of their product rises with the size and complexity of the client’s database systems, and so it is no wonder that Snowflake targets only the largest players on the market. What is even more interesting is that Snowflake operates at very low, or even negative margins [2], which is said to be mostly due to high costs of customer acquisition – after all, it takes incredibly skilled and experienced salespeople to land deals with the big brands they tend to work with. This seems to be an issue, as the cost of customer acquisition looks like typical, startup-like “burning of the cash”, however if we look at the data regarding customer retention we will see that it is a healthy, well run business: Snowflake has an incredible 170% net revenue retention rate with more than 6,000 customers. They customers are extremely happy with their services and even if, hypothetically, the company stopped pursuing any new deals now, they could still be surviving off their existing client base. After all, they manage to simplify an extremely complex problem of data processing and storage on behalf of the companies, saving them millions of dollars.

Time will tell how the data warehousing industry will evolve, but at its current course Snowflake can very well become a monopoly in the market, just like Salesforce is in CRMs or Slack is in mid-tier company communication.

References:

[1] Iqbal Ahmed, Everything You Need To Know About The Cost Of Building A Data Warehouse, 31.03.2021, https://www.astera.com/type/blog/building-a-data-warehouse-cost-estimation/

[2] Snowflake Inc., Snowflake Reports Financial Results for the Fourth Quarter and Full Year of Fiscal 2022, 02.03.2022, https://investors.snowflake.com/news/news-details/2022/Snowflake-Reports-Financial-Results-for-the-Fourth-Quarter-and-Full-Year-of-Fiscal-2022/default.aspx

No ratings yet.

For many years, sports were not really linked to advanced data. Sure, some of the professional leagues were gathering player and team statistics (with one of the frontrunners being NBA, where stats are widely available since 1950s), but it was mostly used for general assessment of player’s quality (e.g. number of points scored per game, etc.). The use of professional data analytics in sport to gain a competitive edge has been initially popularised in baseball, by the Oakland Athletics General Manager Billy Beane, who, in the late 90s and early 00s started applying data analytics to find players, who could be added to his team’s roster and could provide high quality at a cheap price (his persona was later popularised by the 2003 book “Moneyball” and a movie with the same name).

It may seem like football was late to the party. Throughout the years the performance of players and assessment of potential transfers was based on “eye test” done by football scouts. This is, fortunately, changing. More and more football clubs are investing in data science teams to provide better insight into their own players, but also competition and future signings.

When it comes to data-heavy team management, the best case study can be provided by the English football club Brentford FC. In a similar fashion to Oakland Athletics, the club is being run with a limited budget, putting huge attention to data and new technologies. The club has been purchased by former Bank of America VP turned professional gambler, Matthew Benham, who introduced his knowledge of algorithms, statistics and data research to the team management, which allowed him to assemble a great team using very little capital, and improve team so much, that they have progress from the fourth to the first tier of English football. Benham’s data team analyse multiple players all over the world in order to find “hidden gems” and to bring them to London at a cheap price. Some of the players are then sold to other clubs in order to increase the budget and financial capabilities of the club.

Besides using data to improve the management of the entire club, there are multiple initiatives to assess the performance of players, coaches and teams on the pitch. There are multiple companies, such as SciSports, WyScout and Sentient Sports [1] who provide detailed analysis of the quality of players as external entities. The business then sell the data or implementation of their software to clubs who want to improve their approach to advanced analytics. And nowadays, everything can be measured – the sport went from collecting numbers on goals and assists to GPS-based monitoring of player’s position and behaviour on the pitch, heat maps and AI-based algorithms who provide tactical insight based on game videos.

Though it is not only external companies that collaborate with football teams. Many clubs try to develop in-house solutions. Manchester City (together with its partner clubs from Spain, Australia etc.) have employed an ex-Harvard and Yale researcher with background in astrophysics, Laurie Shaw, to lead their data team. German club TSG Hoffenheim [2] has created a tech-heavy, centre for training where real-life match scenarios are generated on screens and the footballers are asked to react to practice their decision making. The club has been reliant on data and analytics for a long time (partially due to their connections to SAP who are their co-owner) which has allowed to to progress through the German football system in just a couple of years.

The use of technology in football is therefore growing, though for now only the elite seem to be implementing it widely. I believe that with lowering prices and growing expertise among club management boards, the implementations will become even more popular, in a growing number of clubs and leagues.

References:

[1] The Register, A short history of data analysis in football, 29.07.2022, https://www.theregister.com/2022/06/29/a_short_history_of_data/

[2] Feng Zhu, Karim R. Makhani et al, TSG Hoffenheim: Football in the Age of Analytics, 08.2015, https://www.hbs.edu/faculty/Pages/item.aspx?num=49569