Why does this matter? Buyers at manufacturing companies face high pressure to cut costs. Supplier sourcing is still very manual and takes up to 45% of a buyer's time. Due to the lack of research time, buyers rarely close deals at optimal conditions. Recently, Covid-19 and material shortages have further accelerated the need for more transparent supply chains.

Isabel came across this problem during her work as an investment analyst. She realized that optimizing procurement – the buying of goods and services – is one of the most important ways for industrial firms to save costs: more than 50% of revenue is spent on COGS (Cost of Goods Sold) [1] and sourcing is reported to be the biggest value-driver in procurement [2]. As an investment analyst she was used to a much higher level of data transparency that is available from tools such as Bloomberg or CapitalIQ. Inspired by this, we set out to build a similar experience for buyers at manufacturers.

After talking to lots of buyers, we found out just how manual current sourcing processes still are: managing existing suppliers in spreadsheets, web research to find new suppliers, calling and visiting suppliers on the ground. Most buyers simply don’t have the time to do a thorough supplier discovery, let alone gather all key information about them to make informed decisions.

Solving this problem requires gathering and analyzing supplier and supply chain data from many sources. Building such a tool represents a massive data challenge: gathering, structuring, and indexing 10B+ parameters of 10M+ suppliers worldwide from 500+ data sources. Our two biggest technical challenges are a) providing a standardized mapping of the supplier’s product portfolios and b) continuously tapping into new data sources which need to be incorporated into our existing pipelines. We use NLP models to extract specific keywords in order to create a list of potential supplier candidates. We match the suppliers to various sources using NER, which allows us to filter down candidates based on specific parameters. The result is a final supplier list with detailed information on company facts, products, financials, and much more.

Based on this data, we run sourcing projects with our customers, which means finding suppliers for a specific product and region. Since our launch in January, we have been able to win several Fortune 500 companies, including BASF and ABB, as customers. Using our software, our customers have been able to reduce time and procurement spend significantly. On a recent sourcing project, we helped ABB find new suppliers for a custom lens made out of a special material and even suggested additional suppliers for a different material with similar properties. For this project, we saved ABB 40% in procurement spend and 3 weeks of manual work.

Our sales process begins with a demo exploring our tool with a subset of the customer's existing suppliers. Customers usually continue with a paid pilot - one sourcing project in a certain region. In order to access their data going forward and book more sourcing projects, customers enter into a yearly SaaS contract. Based on our current pricing, we estimate our addressable market to be at $15.8 B.

You can visit our HN-trial signup page (https://alpas.ai/hn-trial-signup) to create a 48-hours trial account for our software with a demo sourcing project. Please note that we usually do a guided onboarding process with our customers and thus have not optimized our software to be self-serve.

We will gladly answer your questions and are excited to hear your thoughts, ideas, and feedback!

[1] http://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/... and https://www.readyratios.com/sec/ratio/gross-margin/

[2] https://www.mckinsey.com/industries/consumer-packaged-goods/...