1. Rocket Internet’s Modus Operandi


    I’ve talked to many people who’ve worked at a Rocket Internet venture, I’ve had several long discussions with management of Rocket Internet’s portfolio companies, and I’ve also had personal experience dealing with them in various capacity. These are my thoughts on Rocket Internet and how they work. 

    As most of you might know, Rocket Internet was founded by the notorious Samwer brothers. They primarily clone companies that have been proven to be successful in the States and introduce them in new markets (e.g. Amazon clone - Lazada in Southeast Asia, Uber clone - EasyTaxi in developing countries, Groupon clone - CityDeal, HomeJoy clone - Helpling). Here’s a list of all the companies under Rocket Internet’s portfolio. 

    I’ve come to find that Rocket Internet ventures are actually very predictable. Now, if we simplify a majority of their portfolio companies, we’ll be left with only several building blocks.

    1) These ventures have to involve demand and supply, and they must have control over both of them. They’ll control demand through massive advertising expenditure, and supply through the hiring of a large, and typically young and hungry sales force. 

    Take Lazada for example. On the supply side (i.e. items available for sale on Lazada), they have a team of sales people under ridiculous targets to crawl the web, generate leads, call, pitch, and bring onboard vendors, from principals and distributors to resellers. The goal is to increase the assortments available. I’ll describe more on their targets further down below. 

    On the demand side (i.e. users and revenue), extremely large advertising expenditure is the common practice. Lazada is the third largest digital advertiser in Southeast Asia in terms of budget, right behind Jawbone and P&G. That’s right. Lazada’s advertising budget is larger than Unilever. They’re spending approximately USD$5 million / quarter. The same goes to FoodPanda, EasyTaxi, Zalora, Lamido, you name it. Of course, their advertising expenditure differs from one venture to another depending on market size.  

    These allow them to grow recklessly and push out other players if any. 

    2) These ventures don’t do projections based on what they can realistically achieve. They project based on numbers that’ll raise their next big round

    Unlike most ventures, Rocket Internet do projections in a top down manner. They’ll decide what values will allow them to raise their desired round size, lay them out on a spreadsheet and begin hiring and setting targets accordingly. 

    What this results in is insane monthly compounded target increment. It is not uncommon to hear of sales targets, in the Rocket Internet venture sense, increment of 25% every month. As a result, the typical length at which an employee sticks around in a Rocket Internet venture is 4 to 6 months.

    These all make sense if you examine the background of the top management at each venture. Almost always, they’ll be people with prior management consultancy experience (think McKinsey, BCG and to a lesser extent, Bain) and occasionally, private equity. Given the nature of their previous work, they are very handy with spreadsheets and know what is necessary to raise a certain level of desired funding, but not necessarily experienced with operations or running of a healthy organisation. 

    3) These ventures place different teams in different regions as a cost cutting measure

    If the venture is a regional venture, this is almost assured. Lets take the Southeast Asian market as a reference point. Headquarters will usually be in Singapore for the legal structure, QA could be anywhere between Thailand, Vietnam and Philippines for the cost savings in salary and marketing and product development in Malaysia or Indonesia for the available expertise for low cost. 

    This allows them to optimise cost vs available talent throughout the region. That also means that there are very little communication between teams, create dissent, which from what I’ve observed, is their greatest weakness. This is probably also why they’re incapable of innovating beyond cloning.

    4) These ventures share top management, resources

    Now that Rocket Internet ventures have been around for some time, they’ve built up deep execution knowledge. That means, if a certain execution skill is needed by one of the ventures, for example, new market entry, they’ll transfer management from another venture that has done it before to lead the initiative. “Co-founder” title for any Rocket Internet ventures basically mean a transferred manager who is leading a new market entry for a Rocket Internet e-commerce clone. Every country has a “co-founder”. It’s the prime example for an organisation with inflated titles. 

    Despite being an “incubator”, Rocket Internet is essentially a very large organisation with shared resources. Which of course, explains their addiction to marketplace, e-commerce type of business. 


    What these all mean is that if you’re currently running, or is planning to run a marketplace or an e-commerce startup of any sort with demand and supply side problem, with any success, you will be the next target. They have a blueprint that they will not hesitate to copy and paste. Now, if you are running a product company, you’ll most likely be safe from a Rocket Internet style cloning. 


  2. "That’s Not My Job"

    These words could not be more true. I’ve always observed that there are 2 fundamentally different ways of thinking that affect people’s working style vis-a-vis their colleagues.

    1) People who believe that they are hired for their knowledge and should only be doing what they’re good at
    2) People who believe that they have to do, pick-up whatever’s necessary to get the job done

    In a large archaic corporation, 1 is usually the preferred style. Each department are defensive over their own territories. Ego and the need to look good to upper management take precedence over overall organisational goals. This then bleeds into the hiring process, where people are hired for one skill (be it sales, marketing or hacking) and are not expected to learn anything else. As a result, each department is oblivious to the responsibilities of another. For example, in some companies where marketing must have a say in every aesthetic decision, the product management team typically suffers from inconsistent and impossible demands precisely because marketing couldn’t be bothered with learning about the technical development process. To them, their job is “marketing” and everything else is none of their business. 

    However, in a startup environment where speed of execution is of the essence, 1 is absolutely toxic. It’s toxic not only because it’s slow, but also because it affects team morale. The worst thing that can happen to a startup is to be a startup with the speed and morale of a large archaic company and the resilience of a startup.

    In a startup, everyone is responsible to get the job done and get it done fast. Don’t get me wrong, I’m not talking about the “just get it done, I don’t care” type. I’m talking about learning how to reduce friction when passing projects from one person to another. 

    I’ve always made it a point to do that. Lets say i’m working on a project and I need to get the developers involved. The most sensible thing to do in preparation is to read through technical documentations and highlight parts that are required, whatever’s necessary to make the life of the developers easier. If I don’t know how to do it, I’ll learn and if I can’t figure out where to start, I’ll observe this time round and I’ll be sure to contribute more for the next. After all, I’m the one who started it. Developers have their own day-to-day responsibilities too. It’s unrealistic to expect them to understand the requirements AND read through a new set of technical documentation just because I needed it done. 

    In a way, it’s about treating your fellow colleagues with respect. Throwing your weight around might get the project done, but in the long run, it’s going to create a lot of animosity in the company. 


  3. Branding Lessons From Little Creatures Brewery


    Located in Fremantle, the groovy region right outside of Perth, Little Creatures Brewery is one of my favourite place around. The Fremantle brewery is Little Creatures’ very first brewery, and the structure of the building, tall and wide with metal fences on the inside is very much the result of the building’s historical function; a crocodile farm. They’ve come a long way, having expanded to a couple of locations around Australia and is publicly traded on the Australian stock exchange. If you ever stop by, do go on a free tour conducted by Little Creatures’ staff which will introduce you to the various stage of beer brewing, and ends with a tasting session for all the different beers (and ciders) they have on tap. Their selection of food on menu (especially the chili tomato mussels) is not to be missed too. 

    But what is truly exceptional about Little Creatures is how consistent they are at branding. 

    Firstly, the staff. Every customer-facing staff exudes a set of personality that defines Little Creatures. Friendly, hip, groovy and on the ball, with an edgy appearance. There’s something unconventional about them, much like Little Creatures craft beer. Grab any staff at their brewhouse and ask them anything you like about the beer brewing process. You’re in for a lesson. Most, if not all of them, do brew their own beer out of love and passion for it. The consistency in their staff personality is outstanding, which makes me wonder about their hiring process. What’s their hiring process like? And why do these bunch of passionate and knowledgeable people choose to work here? 

    Secondly, the beer brewing. If you went on the tour around the brewery, you’ll noticed that it is unusually clean. As I was told, that’s because at the core of beer brewing is microbiology. This makes it very important to not only ensure a consistent process, but also to maintain a clean and uncontaminated brewery. Which means plenty of cleaning up. You might think that this is commonsensical, but try visiting other breweries and you’re in for a rather shocking experience once you’ve been to Little Creatures. The contrast is startling.

    All in all, the environment, vibe, staff knowledge and tour are exceedingly consistent. I do realize that I’ve been using the word ‘consistent’ over and over again, but that’s because they are. These touchpoints exudes the same positive message about Little Creatures although they function different purposes. The result is customer loyalty and positive brand association.

    If you are ever around Fremantle, I urge you to give Little Creatures an entire day. Go for the tour. And stay for the beer. And the food. And the vibe. And observe. 


  4. Learnings From A Real Estate Project

    I worked on a side project with a couple of my friends awhile back to make it easier for people to look for a place in Singapore. A little backstory. All of the existing real estate listing players seems to be doing a pretty bad job. They provide little to no information and most listings have terrible images (if any). So we thought, wouldn’t it be cool to have one that is actually useful? It’s obvious, isn’t it?

    What we found while working on it shocked us. The project has since been called-off, but it was a great learning experience.

    In Singapore, most people rely on real estate agents to 1) find a place for themselves or 2) find people who are interested in their property. Agents manage their listings, take calls from potential interested parties and handle the related paperwork. As a result, most listings online are managed by real estate agents.

    Here where things get interesting.


    1) You’ll get very little traffic as an individual listing your own property

    The largest and arguably most successful real estate listing site in Singapore is PropertyGuru. You can search for properties, and search result will be arranged in the order that is most relevant to your search term. Or is it? 

    Turns out, you can influence the search result. Check out “Priority Listing” http://www.propertyguru.com.sg/agents 

    There are altogether about 12,000 real estate agents on PropertyGuru, each, competing for eyeballs, each paying large amount of money to ensure that they appear as high as possible on PropertyGuru search. It’s their career. They will willingly pay the sum, and spread the cost out to  each of their sales. 

    Chances are, your independent listing won’t be given any priority. 


    2) Real estate agents want you to have as little information as possible

    Real estate agents want you to meet them. It’s easier to convince you that the property is exactly what you’re looking for in a face-to-face environment. It also makes them look good to landlords. Getting people to view properties signal to landlords that they’re doing their job. 

    But the more information you can get online, the less likely you are to meet with them. You already have all you need for comparison, why bother going down for viewing? So housing agents provide as little information as possible. That means, no pictures and if there is one, it’s definitely not a picture of the house. Probably the pool, or the balcony. No descriptive information about the house. You get the idea. 

    But what you’ll definitely find is the phone number of the housing agent. 


    3) Many properties listed online are fake

    Housing agents earn through agent fees. They can earn agent fees from both landlords, and home-seekers.

    Note: I am unable to verify anything from here on in point 3, but I’m relaying these information the way I was told by housing agents that I have spoken to. 

    What I was told was, if a housing agent advertise an available place, and home-seekers contact the agent about the advertised place, the agent cannot, by law, collect agent fees. So in fact, housing agents are only able to collect agent fees if they do the matching, and they can’t collect agent fees if contacted directly regarding a property. Of course, regardless of outcome, they’ll always be able to collect agent fees from the landlord since the landlord engaged them to do the job. 

    Now, if they were to post available properties online, and you, as an interested renter were to contact them about the property, they won’t be able to double-dip. You are contacting them directly.

    Here’s where fake properties come in. There is a loophole. 

    Scenario 1. You saw a property advertised online, and you’re highly interested. You contact the agent, and the agent tells you that he/she’s sorry, the property is not available anymore. However, another property nearby that is very similar is available! Would you be interested? 

    In this scenario 1, the agent can collect agent fees. He/she’s matching you to another property. 

    Scenario 2. This is where the agents are more entrepreneurial and work in teams to share sales leads. You saw a property, you mentioned you’re interested, the property is not available anymore. Same story. Now the agent is very sorry, but didn’t offer another property. However, in the coming days, you get calls after calls from housing agents who know that you’ve been looking at nearby properties. Again, the agent do the matching, therefore they can collect agent fees. 



    What I’ve found is that ideas never matter. There are always something about an idea, something troubling about it that have to be overcome, and that something is usually very difficult to tackle. 

    But that doesn’t mean we shouldn’t try. Sometimes, naivety is our best asset. It lets us learn things at an accelerated pace, things that will help us execute in the long run, all without the reservation that we would have if we knew how difficult things are going to be. 


  5. Facebook Likes and Twitter Followers are Vanity Metrics

    TL;DR - Don’t focus on getting likes on Facebook, or followers on Twitter. That is not a marketing strategy. They’re useless if they don’t serve a purpose. And by purpose I mean increase revenue, increase users and/or increase usage. Instead, what you should be doing is looking at driving traffic through a carefully optimized ad campaign if you’re going to spend money and effort at all. 


    There are a lot of “social media experts”[1] out there who will tell you how important it is to increase your company’s social media presence. 

    “You need to get more followers on Twitter” - they say
    “You need more likes on Facebook” - they say

    If you’re lucky, you might even catch them uttering the word “viral” sometimes. The funny thing is, none of them will ever mention why that is important (it’s not).

    Try asking them why the next time you have the chance. Ask them. Go ahead. Chances are, their answers will unanimously be because you’ll have a larger audience to market to. Free marketing for life, right?  


    Let me tell you why. Just to be clear, none of these have anything to do with ads. Just plain social media following. Ads serve a difference purpose and they can serve you pretty well if you track and optimize them well (I’ll write about this in another post). And, if you’re going to do anything with Facebook, make sure you allocate a portion of your budget for Facebook marketing as well. 


    Facebook Page Likes

    If you’re still trying to get Facebook Page likes, or optimizing content for Page Post likes, you need to stop. 

    Your FB page likes are not as useful as you think it is. If you’ve got 100,000 people liking your page, you should count yourself lucky if 20,000 of those people saw your carefully crafted, meticulously timed posts[2]. Facebook throttles your posts. Here’s another one by TechCrunch

    The reason is very simple. As Facebook users get more and more connected with each other, their News Feed stream get more and more crowded with status updates. This presents an interesting problem. If all everything is displayed in the News Feed, most of the content will turn out to be irrelevant to Facebook users and this in turn, cause massive user experience problem. 

    To counter that, Facebook programmatically decides what is relevant and what is not relevant to the users based on past behavior. This means that all posts are not created equal. 

    In short, your posts don’t go out to everyone. On a busy day, expect even lesser people to see them[3].

    This is where ads come in. Now, if you really want people to see the page posts,  Promoted Post is your answer. So my question is, why bother putting in money to increase likes in the first place? 

    In fact, I’ll like to go one step further. Why focus on a Facebook Page? Will anything you post directly generate revenue, convert new users, or drive usage? If yes, automate the process. That way, you have a workflow that will help you save time, and when you really want to do a push, you can do it through Facebook ads. 


    Twitter Followers

    Twitter followers are a lot more useful. You can actually get real users (that also happen to become followers) without spending a single dollar. Austen Allred has a very comprehensive guide to getting users on Twitter. Couple that with proper tracking, you’ll have a pretty clear understanding as to what attract users and what doesn’t.

    Having said that, making increasing Twitter followers a goal is bad for your company. You’ll start to optimize for retweets and favorites, and before you know it, you’re spending all the time you could have spent more productively, trying to come up with things that you can get retweets from[4]. Worse, these tweets are in no way helpful to increase your revenue, user base, or the usage of your product. Soon, your followers will have no idea what your company do from the tweets. 




    Followers and subscribers are vanity metric. They make it easier for you to brag, but they’re not very useful apart from reputation and branding for your company. What you should do instead is putting the effort and money into what can directly increase your metric. Users, usage, revenue. 

    And the easiest way to do that is through ads, although you do have to be careful and spend prudently[5]. The good news is, if you really like page likes and followers, ad spending on Facebook and Twitter typically have the spillover effect of increasing those as well. So fret not. 


    1) A paraphrased quote from an acquaintance. “Every high school and college student who spends way too much time on Facebook claims to be a social media expert”

    2) This is also from observation of Spuul’s own Facebook page data.

    3) Facebook needs to balance what appears on a user’s News Feed. Priority are given in this order; relevant posts by friends, revenue generating ads and free page post.

    4) You are what you measure. I’ve personally seen optimization for retweets and favorites end up making tweets totally unrelated to the business. The scary thing is, you won’t even know you’re doing that without someone pointing it out to you

    5) Many people who claimed that Facebook ads are way costlier than Google ads managed their Facebook ad campaigns wrongly or have used Facebook ads for the wrong reasons. E.g. a page post with an image (uploaded onto Facebook) as the centerpiece of the ad, while the aim of the ad is to drive traffic to the site.


  6. jQuery Not Loaded In Rails 4 App And The Solution

    I’ve just started learning Ruby on Rails and if you’re like me, you’ve probably forked a couple of test projects/exercises to learn by getting your hands dirty. 

    I went straight to Ruby 2 and Rails 4 knowing that there might be incompatibilities here and there when forking older projects. Lo and behold, there are incompatibilities indeed.

    The main issue that I’ve faced so far (and took me awhile to find the solution to) is jQuery not being loaded on projects. After searching all over StackOverflow, there aren’t any solution that actually solved my problem.

    Most commonly suggested solutions:
    - check /app/assets/application/javascript.js and make sure standard code generated (it is)
    - check Gemfile to make sure that gem ‘jquery-rails’ is included (it is)

    The thing that really fixed it for me was only mentioned once (as far as I can find) on StackOverflow. Older version of Sprockets stopped working with Ruby 2

    By just updating sprockets to 2.10.0 (the latest version at the time of this post) and bundle install, the problem is fixed.