Why are there no massive Chinese SaaS companies?

The 🇨🇳 SaaS series #️⃣ 1; Bonus: The F theory of startup evolution

Welcome to all the new subscribers! I write about tech in China from the perspective of a 🐢 VC with a new update every Wednesday. If you like the content, please share with a friend!

As a longtime SaaS investor, a question that has been close to my heart is - where are all the giant Chinese SaaS companies? I define Software-as-a-Service (SaaS) as both a licensing and a delivery model, meaning the software is charged on a recurring basis and hosted in a multi-tenant cloud. Even though not all software is SaaS, increasingly the two words have become interchangeable in the West among the new generation of software.

As a baseline, here's a chart of the 20 biggest US tech companies (excluding hardware) by market cap on 29th August 2020 - enterprise companies account for 55% by numbers and 40% by value. I know the definitions of both 'software', and 'consumer' versus 'enterprise' are contentious (Apple is charging for services! Amazon has AWS!), but let’s go with this for now.

And most of the enterprise companies are also SaaS or very close to it.

In contrast - here's a chart of the 20 biggest Chinese tech companies (excluding hardware) by market cap. Enterprise companies account for 30% by numbers and 3% by value and no SaaS companies amongst them.

The SaaS companies aren't hiding in the private market either - here's the latest Chinese unicorn list from CB Insights (Ant Financial is still a subsidiary of Alibaba, so wouldn't be on here), I've capped it to the top 20. Still, the rest of the list follows similarly.

So to answer my initial question of “Where are all the giant Chinese SaaS companies?” - they don't exist yet.

This conclusion is also in line with the macro data on IT expenditure; the chart below is from a Bain report where we can see China lags behind the developed world on investment in IT relative to its size.

I've yet to find reliable estimates for the SaaS market in China, estimates varying between $3.7bn - $6bn in 2019, which is still less than 6% of the total world SaaS market size.

"The enterprise software market in China is lagging the US by at least ten years. It has a long way to go in terms of market acceptance and willingness to pay" - Eric Ye, Partner of Eminence Ventures 

So the follow-up killer question: Why doesn't China have massive SaaS companies?

I have some educated speculation. 

After much secondary and some primary research, the following are emerging as the key demand-side and supply-side factors:

For Chinese SMBs, who account for ~60% of China's production GDP, adopting SaaS is difficult as it's not a learnt reward loop to invest in software of any kind. Historically, cheap labour in China has meant that manual execution is still very feasible for most tasks, so automating or documenting through software is not a frequent behaviour. Coupled with this, Chinese SMBs generally don't have enough cash reserves to invest in upgrades systematically. When they do have the investment capacity, the prevalence of software piracy in the '00s has led to a general lack of desire to pay for software, and many do not often see value in utilising software.

This mindset makes SaaS adoption in China particularly hard as the well-capitalised mid-sized firms are traditionally the landing point for SaaS companies in both the US and Europe. Without a firm client base to build upon, a lot of SaaS companies end up having to tackle the daunting enterprise customers straight away.

For the Chinese enterprise companies, paying for software is not as much of an issue. However, they face a lot of problems with cloud adoption. Below is an excerpt from a Mckinsey report on cloud adoption in China:

'There's the cost and difficulty of migration, which 66 percent of respondents identified as a major barrier to cloud usage. Their concerns are justified, since most Chinese companies have less advanced technology stacks, with most lacking a standard virtualised or containerised workload. These factors complicate all migrations, but the problems tend to be most severe when using a public cloud provider. On the cost side, Chinese companies face big expenses during cloud migration because they must typically build or create most hardware and software necessary for the shift. That means public-cloud vendors in China can't claim that their services will reduce IT costs, at least over the first few years of migration, as they do in other countries.'

Even if they get over migrating to the cloud, there's still a lot of concerns around the stability and securing of a public cloud (not all of which are unwarranted given the nascency of the Chinese cloud providers), with most opting for a private cloud or hybrid cloud deployment.

Both Western and Chinese enterprises want customised software in return for their high ACVs. Most Chinese SaaS companies have difficulty with their customer’s convoluted internal company structure, processes and workflows (derived from what I assume is a mixture of top-down management and lack of project management best practices). This not only makes customisation highly time-consuming but also doesn't allow for much re-use of the work for other clients. The typical SaaS startups dilemma is "should we customise and earn actual cash or focus only on the product roadmap features aka features everyone can use?" The general advice is some variant of "only take the customisation that adds to the product roadmap". However, I feel like China has an extreme case of no overlap between customisation features and standardised features list for most companies. Many have complained that most Chinese enterprise SaaS companies are IT consulting services in disguise.

Other hindrances to SaaS adoption such as wanting to build in-house, distrust of public cloud, even lack of talent in scaling Chinese SaaS companies are reflective of a nascent market for SaaS, and not a mark against the Chinese market in particular. Based on observations of tech development, where the wave of enterprise startups tend to follow consumer startups (more on this in the bonus section), I'd put the China, Europe and US tech ecosystem into the following positions:

While these adoption hurdles are complicated, the progress of western startups has shown that they will get solved with time. VCs who have traditionally focused on consumer startups as that hyper-growth model allowed quicker returns to investment (or death of investment) have started to look for new opportunities. I've seen the sentiment of 'post-2015, the consumer startups are yielding diminishing returns to capital' passed around as fact on a lot of Chinese internets. VCs have also started to focus more on the enterprise software segment and will hopefully have more patience for the longer timeline that B2B startups take to reach scale.

A counter to my answer of no big Chinese SaaS companies - there are, they are just inside the mega consumer companies of Alibaba, Tencent and others. DingTalk (launched by Alibaba in 2015) is an all-in-one mobile workplace. It enables text, photo, voice and video communication, workflow management and collaboration as well as more Orwellian features. Tencent and Bytedance also have their competitive offerings in the form of WeChat Work and Lark. Even Huawei has got into the fray with Welink and potentially Pinduoduo as well. Collaboration offerings aside, there are similar enterprise software offerings for manufacturing and logistics from these big players as well. These consumer tech companies have the resources and workforce to move into the enterprise space, and they make it very difficult for new entrants to flourish. How? Currently, DingTalk has 300 million users in 15 million enterprises (helped by COVID), but the service is free. That's hard for most startups to compete with (esp when they often don't have a large treasure chest of VC money to rely on), and the reinforced learnt behaviour of not paying for software is also problematic.

In my previous newsletter, I mentioned that Chinese consumer tech companies often start with the product philosophy of 'owning the user'. I wonder whether that perspective works in the enterprise world. A lot of Chinese SaaS offerings are incredibly comprehensive in terms of functions but misses a killer feature that they are known for. In wanting to be all things to all people, these SaaS offerings often seem like wholesale pastiches of successful western startups without accounting for the Chinese market's nuances. In terms of designing for painkillers and managing customer success effectively, the SaaS startups and consumer corporates are both still trying to figure it out. The lack of a quick feedback loop from a buyer base doesn't help either.

The state of play of Chinese SaaS does seem a bit dire, though a lot of the influences mentioned are waning (such the end of demographic dividend will mean increasing labour costs in the future). At the same time, new trigger points have emerged (COVID has accelerated the adoption of cloud and remote working) and opportunities have arisen (the government have been making a big push for cloud adoption). Agora - a Chinese SaaS company that enables developers to add HD interactive broadcast, voice, and video recently IPO’ed and has a current market cap of $4bn - a sign of things yet to come I’m sure.

This is the first part of a series on Chinese SaaS , and I'll follow-up on the turning points and potential opportunities in future posts, stay tuned! In the meantime thoughts, disagreements and feedback super welcome!

(Bonus after the jump but do share if you enjoyed the article so far 🙇🏻‍♀️)

Share Chinese Characteristics


We’ve mentioned this pattern above already, where each type of companies come in succession of the one before.

As to why this general development pattern exists, here is where I'd like to introduce my theory of "F*** this" on startup evolution (F theory henceforth). As conventional startup theory generally proposes that successful startup addresses a key need, my F theory proposes that each successive wave of startups is created through the frustrations experienced in the previous tech waves.

The first wave always starts with consumer tech companies - partly informed by the frustrations founders experienced in their life as a consumer (therefore a lower barrier to having a workable idea), partly it's a sector that gets traction very quickly. So a bunch of consumer tech companies spring up. They employ a lot of people, and everyone is super keen on using tech to solve every problem in the company, and pretty soon, people in those companies start getting fed-up with their internal systems. Why is recruiting so bad that we don't know how many people we're interviewing? Why are we still using Excel to visualise all our data? Which department is doing knowledge management again, what do you mean all we have is a word doc? All of this piles up until a person one day goes 'F*** this, I'm doing this myself, and I'm going to make a company out of it', or some lengthy variant of that.

Once enough people within consumer companies start doing this (and get financed by VCs who love that narrative), a new wave of enterprise tech companies springs up. They employ a lot of people, and everyone is super keen on using the latest enterprise tech in the organisation, and pretty soon everyone starts getting fed up with various functions and processes - why do we have to build <insert> from scratch each time? Why aren't we monitoring the meta-data of xyz? How are we managing these two systems to talk to each other again? All of this piles up until a person one day goes 'F*** this, I'm doing this myself and blah blah blah'. Lo and behold, a new wave of meta-enterprise companies (which are companies who exist to serve the needs of other tech companies or tech departments within companies)** springs up.

I subscribe to the creation-through-rage school of thought.

Postscript: The evolution pattern also indicates the growing market awareness and demand for these offerings. Once the consumer starts to experience digitised retail and communication, they are more open to tech offerings in their workplace. Once most industries start to undergo' digital transformation' with enterprise software or begin to digitise themselves to keep up, more demand will come for the developer tooling and infrastructure tech.

*Amazon is still mostly a consumer company - look at their revenue breakdown….I don’t want to talk about their profit breakdown.

**Some examples: I would classify Salesforce as an enterprise company and Twilio as a meta-enterprise company, while a company can use Salesforce without developers (at least theoretically), a company can’t use Twilio without developers.

Amazon shows how this can happen within a company's product offering. They started with Amazon marketplace, then started branching into enterprise offerings with AWS and eventually meta-enterprise offerings such as AWS Lambda.