1. Only Platformization Lets You Go Further

As long as you respond to the market a little faster than your competitors and live a little longer, you can win more room to survive. A brilliant business model or market opportunity is more like magic: it can grow a company fast, but it is not something human effort can control. What we can do is polish our tools and platforms, so that when the wind rises we can change with it — with a little mastery, we can ride it straight up.

Platform construction is a long-cycle, high-return investment. A platform grows along with the company, and a big company is bound to have a good supporting platform.

Most small companies do not survive 3-5 years. Small companies will not invest heavily in building platforms, but they are very happy to try fresh technologies. New technology means not mature enough and not stable enough, but it buys extra competitiveness. For example, adopting a new release system that supports 10 versions in canary at the same time greatly speeds up product iteration and improves the product’s competitiveness in the market.

The new platform does bring benefits to a small company, but as the company grows, such off-the-shelf platforms gradually fail to meet business needs: we have to keep investing manpower to maintain someone else’s platform, or be willing to pay for it.

Note that the marginal return of a platform supporting business growth declines over the long term. The reason is that in the short term you benefit from your own scale growth, but as people move through the hiring market, that gradually erases the advantage over competitors and reduces the return the platform brings. When the platform loses its advanced edge and the excess return it generates no longer covers the cost spent on it, the company starts looking for a new platform.

To a certain extent, how advanced a platform is represents how advanced the company’s means of production are. Leading platform services make a company more competitive in the market.

2. The Height of Your Understanding Determines How Advanced Your Platform Is

What exactly are we talking about when we talk about a platform here? I think a platform is a kind of service-ization of knowledge. That knowledge comes from accumulated understanding of some domain — it can be experience, it can be general know-how, it can be a comprehensive solution.

Building a platform without a high level of understanding is like making a submachine gun: the user’s ignorance and ill intent can cause enormous damage. So before building a platform, we must raise our understanding of the domain and abstract the relevant concepts into a logic that fits our own business.

Accumulating understanding is a hard thing, and abstracting and generalizing it is even more of a challenge. The frontline team faces the business scenarios directly and has rich experience, but is busy with trivia and has no time to think; the supporting team has tools and methodology, but finds it hard to get down to the frontline and fully understand the business scenarios. Collaboration between these two teams is costly, but only after this kind of experience propagation and the transformation and distillation of understanding is achieved does a platform have any chance of landing.

3. Move Toward Service-ization, Away from Process

Standardization is the prerequisite for service-ization. Only standardization lets a small company grow into a big one. Personalized, customized things are always niche; universal, common things have enough market space. The process of a company growing is in fact the process of pushing from the local to the global, from the niche to the masses. And to make more people accept it, secularization and scale are indispensable steps.

What is served is people. A platform does not face machines that follow orders, but people who make mistakes, overlook things, and have emotions. A platform should not make the user go to A first, then B, and finally C to reach a goal; it should chain A-B-C together and let the user authorize.

So with service-ization here, what I want to emphasize is that the user only needs to authorize and does not need to know the internal process details. It is like withdrawing money from a bank: you only need to enter your password and you get the money; the bank does not hand you an operations manual and have you follow the instructions in it.

But the reality is that once a platform action crosses teams, you must read documentation and understand the process to complete it. That experience is terrible: the product form is too mechanical and lacks a service mindset.

The root cause of the problem is that multiple teams lack a shared vision — a process-oriented delivery mindset. What we should understand is that merely developing a feature does not bring any user value; value only comes once the feature is live and actually being used. So we should abandon the assembly-line way of working; passing work along one step at a time is not only inefficient, it also fails to ignite the team’s enthusiasm.