Comment by juliansimioni | Hacker News Reader

Comment by juliansimioni | Hacker News Reader

I moved to Berlin for a while in 2014, and so I ended up interviewing at a couple Rocket Internet companies.

I'm by no means an expert in any of this but the impression I took away was that while they did some really impressive things, all their companies had a fatal flaw.

The impressive things were:

- They had a common tech stack that while no means elegant was at least effective at getting to decent scale quickly

- They clearly had marketing and growth talent to also scale up businesses quickly

- They had expertise in specific geographies that are often difficult like South America, Europe, and China (these are difficult at least from the perspective of an American :) )

But the fatal flaw all their companies have, by design, is that they are really good at cloning a product and not good at actually creating one. Every company was clearly designed to be acquired before that problem became insurmountable.

If the acquisition didn't happen, turning a several hundred person organization into one good at figuring out what changes to make to a product to make it successful is...messy.

I worked at Foodpanda, which was one of the Rocket Internet flagship companies for a time. Foodpanda was a mess, and the people from Rocket Internet were saying it was 10x better than Rocket Internet. Which I can believe since Foodpanda outsourced work to Rocket Internet because Rocket Internet said so. They literally assigned me to make sure they didn't screw the project up, they wouldn't take any feedback really. I ended up not caring about the project because it wasn't actually within my offical responsibilites. At the end of the project I was included in the email threads between the two CTOs where the Foodpanda CTO repeatedly pointed out that they were warned about all the shortcoming in the project way before the end and that they needed fixed. Rocket Internet responded saying that was a lot of work and they would need developer resources from Foodpanda. Me and one of the guys I worked with at Foodpanda decided a way to tell if a developer was worthwhile in Berlin was how long they stayed at rocket. Majority in Berlin would have worked for them at some point, so working there wasn't a sign they were bad. But if they stayed there for more than a year, that was.

From my experience Foodpanda's biggest problem was tech. They literally had hundreds of servers be knocked off line because one redis server with session data went down. This was considered acceptable within the company. There was a culture of failure. And the business side had these issues too. There was a write up in an Indian newspaper where they basically got tore to shreads for all their failures publically. Turned out half of the company didn't realise how bad things were.

Foodpanda was Rocket's attempt at buying up competitors and dominating India. They literally bought everyone and still lost the war to dominate food ordering in India.

First of all I'll admit I was trying to err on the side of generosity when describing Rocket. Maybe that's a mistake.

They definitely seemed to have a culture of well...internal chaos. But that jives with their "copy fast" mentality. It's worked at least a couple times, right?

> It's worked at least a couple times, right?

Yea, HelloFresh seems to be doing well and Zalando dominates in fashion. There are probably a few others doing well too.

Delivery Hero too, they just entered the DAX

I am not too sure about them. They sold the German companies and pulled Foodora out of a few countries.

Look at their global results, Europe is largely irrelevant compared to APAC and MENA. Germany exit was loss of prestige losing home market but also not a major loss of revenue compared to gains elsewhere in the world.

Globally, it's mostly Foodpanda that was doing soo poorly they had to be sold to Delivery Hero, no?

All Rocket ventures were designed to be sold. Most of them did poorly. There is a cost/benefit ratio at which it doesn't make sense to dump money on the business, but the business itself still has value, so selling is a good strategy to cut your losses and/or make some profit.

I think you missed the fact that Foodpanda and Delivery Hero were both Rocket companies. It was that Foodpanda was doing soo poorly they gave it to Delivery Hero which was for Europe. Overall, Delivery Hero/Foodpanda are pulling out of markets repeatedly. And not small markets either.

Also, Foodpanda's plan when I was there was IPO.

DeliveryHero isn't a Rocket company. They invested heavily in it at some point in time - and they seem to still hold a position in the company, but DeliveryHero started outside of Rocket. Foodpanda was always pretty much a "rocket venture".

When rocket invests heavily, it becomes a rocket venture. That's how VC pretty much works.

Zalando were publicly hung out to dry by Rocket Internet as they were sure they would fail. After the turn around they are now the poster company of Rocket Internet despite Rocket having done nothing to help them.

Zalando seems really nice. I have no connections at all, but they really won me as a loyal customer because I really prefer the experience over all alternatives (especially including Amazon and local stores for everything except really pricy clothing).

More importantly though, I am really impressed and surprised by some of the research coming out of Zalando. Flair[1] has some great ideas (especially at the time of release, when there weren't pretrained huggingface models for everything to build upon) and a really well-written paper. Colleagues also have had a good experience using the software and achieved very results with adapted NED tasks.

[1] https://research.zalando.com/welcome/mission/research-projec...

> They literally had hundreds of servers Woow. I remember when you could write the number of servers for India with few binary digits :)

I was familiarized with Foodpanda in the early days, and their first months with dev in Berlin. I was also quite familiarized with Rocket (worked there for a while on their core platforms - both ecommerce and marketplace), and while my opinion about their design (on ecommerce) is far from good,their code and procedures were at a level of maturity far more advanced than Foodpanda.

That is not necessarily a bad thing, in startups of this kind you need to be able to refocus priorities and quickly test new markets. With this comes pressure to implement changes and twisting the product to fit a different goal thought on the initial design. In the end, you usually work a PoC as a product for some years until you have a complete grasp of where your business fits. Some challenges may look unexpected (such as the street naming and numbering somewhere in South America, maybe Colombia?), some challenges can be a bit more predictable (such as an appropriate redis configuration), but these growth pains are expected. One way I've seen working quite well to avoid that feeling of "drifting around" product teams have during this trial periods is to have clear, in-person communication from the top down, and include the product teams in the refocus process, so everyone is on the same page, and the business can decide also based on other factors such as cost of implementation and time to market. And I've seen this on a Rocket venture :)

Out of interest, what was their tech stack based on?

I worked there for a while. They used PHP and had a sort of framework called Alice and Bob which I think was built in Symphony.

It was designed specifically to allow quickly rolling out to many different countries each having possibly different business logic.

Bob & Alice is ZF1

Reply on news.ycombinator.com