Why Mapudo failed
A retrospective on the shutdown of a B2B online marketplace in the metals distribution industry
After five exciting years, numerous digital transactions, and a great team effort we have shut down Mapudo on June 12, 2019. Since then we have been approached by many asking us to share the reasons for this decision as well as our key lessons learned at Mapudo. We therefore decided to write a short retrospective elaborating on the factors that might have particularly contributed to the “failure” of Mapudo. We hope the insights prove meaningful for those who pursue a similar vision or are active in related fields.
For those, that are not familiar with Mapudo: Mapudo was an online marketplace for metals, facilitating digital transactions between metal distributors and metal-processing companies. The company was founded in May 2014 in Dusseldorf, Germany, and has been funded by four very supportive investors. While the first version of the platform followed a shop logic, we entirely pivoted the model in the summer of 2018 towards a configurator-based offer comparison model to better reflect the typical procurement processes in the industry (see previous article on achieving product-market fit for more details). We launched the new configurator in November 2018 and managed to improve our conversion rate (visit to buyer) by more than 4x (see graph) and thereby to reduce our customer acquisition cost to less than € 50 until May 2019. Nevertheless, we did not manage to secure enough funding to continue implementing our platform vision. What are the reasons for this?
Reasons for Mapudo’s failure
While it is hard to identify one reason for the “failure” of a company, in the case of Mapudo the core challenge was the too limited frequency of usage by target customers, metal-processing companies. Despite the significant improvement of our customer acquisition cost, we still needed multiple transactions to start earning money with a customer. In theory, this should not have been a problem, as metal-processing companies typically purchase metals multiple times per month. However, in practice the frequency of usage at Mapudo looked different: Metal-processing companies were very happy to use Mapudo for secondary needs (special products, products they had no supplier for, etc.), but weren’t willing to switch their more frequent primary needs from offline channels to Mapudo. There are a few factors that might have particularly contributed to this limited frequency of usage, which we will outline in the following:
1. Low margins in the industry
A digital platform such as Mapudo has the potential to make the order process in the metals distribution industry a lot more efficient, but at the same time adds an additional player to the value chain. As the industry is generally characterized by rather low margins, it is difficult for a platform to earn a substantial transaction fee and at the same time incentivize suppliers to offer competitive prices transparently on a public online platform. At Mapudo, we earned a single-digit transaction fee, but the convenience of the platform for suppliers (i.e. receiving orders without manually calculating an offer) did not translate into prices attractive enough for target customers to rely on Mapudo for primary needs.
2. Stickiness of existing processes
Existing sales processes in the metals distribution industry might be perceived as inefficient from an outside perspective, but they work and are in place for decades. Fundamentally changing such sticky processes on the one hand requires lots of education and on the other hand very compelling unique selling propositions (USPs). While target customers regularly confirmed that they loved the convenience of our configurator (e.g. with exceptionally high Net Promoter Scores), this efficiency advantage alone couldn’t convince them to entirely change their behavior and processes. More competitive prices could have provided a direct cost advantage to target customers and convinced them to also source primary needs at Mapudo. However, based on the low margins and the additional transaction fee of Mapudo, this was hard to realize.
3. Close customer relationships
New customer acquisition based on traditional sales processes has a challenging risk-return profile for metals distributors: It can be very expensive and at the same time yields unclear results. Most metals distributors therefore have a high share of repeat customers and close relationships with these customers. While it was never our intent to replace these relationships, a platform such as Mapudo eliminates at least the personal interaction during a transaction. In feedback calls with our customers we frequently learned that they were very happy with their set of suppliers for their primary needs and liked the interaction with those. Suppliers typically did a great job in serving those customers on a reliable and continuous base, which even allowed them to charge a price premium on certain orders. We were able to demonstrate the potential for less expensive new customer acquisition based on online marketing and target customers’ secondary needs, but the strong relationships with suppliers might have prevented such customers from also using Mapudo for primary needs.
4. Late pivot
The pivot of our platform from a shop- towards a configurator-based customer journey was certainly the right decision, as the significant conversion rate improvement indicates. However, it took too much time and money (incl. the negative implications on our cap table) until we were able to launch the configurator. While the vision for a configurator-based offer comparison model existed early on, it took us more than two years after the initial go-live of the shop-based platform until we were able to launch the configurator. Offer comparison models depend on highly standardized and comparable offers (e.g. in terms of product data and price elements like material price, delivery cost, packaging, etc.). In the metals industry product article numbers and EAN codes are seldom used in daily business. Hence, products need to be identified and matched across suppliers entirely based on their attributes, which requires a standardized product database. At Mapudo, we have spent a significant amount of time and resources on building the first scalable product data base of the metals industry — in combination with the industry specific algorithms of the configurator, Mapudo’s core competitive advantage. Looking back, we could have achieved this faster by focusing on one specific and clearly addressable market niche even earlier.
All in all, these factors translated into a too low frequency of usage of target customers, which — despite improving unit economics — didn’t allow us to raise another financing round in May 2019. We unfortunately ran out of funding to follow through on our clear roadmap towards a higher frequency of usage of target customers. In consequence, we had to shut down Mapudo. We still firmly believe that a digital platform can add substantial value in the metals distribution industry, as the example of Zhaogang in China impressively demonstrates. However, we are also convinced that building the leading platform in the western metals distribution industry not only requires a strong team with a powerful combination of industry experience and digital know-how but also significant time and funding.
We hope that this article provides meaningful insights for others in the field. Feel free to reach out to us in case you have any questions or want to share any thoughts, additions, and critical remarks.
Finally, we would like to take this opportunity to thank our team, our investors, our suppliers, as well as all other partners for the continuous support over years on this exciting journey. We are obviously very sad to not be able to continue together on our clear path, but are grateful for the opportunity of having worked with such an amazing group of people in such an interesting market environment. We hope to stay in touch in the future. All the best, Niklas, Markus & Christian