Thomvest Research: Understand the Chinese Mobile Advertising Ecosystem

by Nima Wedlake on December 1, 2016

It’s no secret that the Chinese advertising market – much like the broader Chinese economy – is large and growing rapidly. As investors in advertising technology, we’ve seen many of our portfolio companies rapidly scale their businesses in China over the last several years.

In order to better understand the Chinese advertising market we kicked off a research project earlier this year, which included a trip to China in July. Our goal with this research is to grasp the many nuances of advertising technology in China – what role does programmatic play, who are the key vendors in the space, what challenges do these vendors face, and how do we expect the market to evolve over the next several years? This report is a culmination of those efforts.

The full presentation is available on Slideshare, and below we’ve included some highlights from our research:

Fueled by mobile adoption, the advertising ecosystem in China is flourishing


There are more than 620 million Chinese consumers with access to internet-connected mobile devices. These consumers have come to rely on mobile devices for commerce, communication & entertainment – Chinese mobile users between ages 16 and 45 spend nearly 40 minutes more per day on their devices than consumers in other countries.

As such, advertising spend on mobile devices has ballooned over the last several years, from $8B USD in 2014 to more than $27B USD in 2016. China is far and away the fastest growing digital advertising market globally, by both impression volume & ad spend.

However, most digital ad spend is being captured by a small set of premium publishers


As Chinese consumers have come to embrace mobile connectivity, we’ve seen the rise of three dominant technology companies – Baidu, Alibaba & Tencent (collectively known as “BAT”). Through both innovation & acquisition, these companies have developed full-stack consumer experiences, including messaging, commerce, gaming, entertainment, and payments. Chinese consumers are constantly using apps owned and operating by either Baidu, Alibaba, or Tencent – 71% of time spent on mobile devices is within a BAT-owned application.

Because these companies capture so much consumer attention, they’ve become advertisers’ primary destination for deploying ad dollars. According to eMarketer, Baidu, Alibaba and Tencent will nearly 73% of China’s mobile internet ad market in 2016.

Cross-border M&A is driving competition & innovation in the advertising sector


In spite of BAT’s dominance, the broader advertising ecosystem continues to thrive, evidenced by the uptick in M&A over the last year. In particular, we’ve seen a number of acquisitions of US and Europe-based companies by Chinese buyers, including the $1.4B acquisition of AppLovin by Orient Hontai Capital & the $900M acquisition of by Miteno Communication Technology. While these acquisitions are often driven by financial arbitrage opportunities, they are also a sign of the growing role of digital advertising in China. We expect these acquisitions to continue into 2017, especially among western companies with a strong presence in the China market.

China Mobile Advertising Landscape Report (Thomvest Ventures) from Thomvest Ventures

Shopping Habits of Smartphone Users & Implications for Advertisers

by Daphne Ewing-Chow on October 26, 2016

In a recent research report, we shared our findings around the rise of mobile device usage and its effect on the advertising ecosystem. We found that adults in the U.S. spend more than three hours on mobile devices each day – playing games, connecting with friends and consuming media. As consumers have shifted to mobile, so have advertisers: ad spend on mobile devices is expected to exceed $40B in 2016. in this post, we dig deeper into the psyche of consumers on mobile in order to better understand its implications on e-commerce.

The Monetate Ecommerce Quarterly provides purchase conversion rates across several device types. The chart below shows that buyers are increasingly more likely to purchase items on tablets; however, the story is quite different for mobile devices which convert at about one-third of the rate of desktop computers.


Statistics from eMarketer’s US Mobile Time and Activities Stat Pack reveal that smartphones are more of a browse or research platform rather than a buy platform, and that consumers tend to use smartphones in tandem with tablets and other larger screen devices when making purchase decisions. The eMarketer study shows that although 8 out of 10 US Smartphone users shop on their mobile devices, less than half actually buy something.


This picture is made clearer when paired with data from Signal around multiscreen shopping behaviors. According to Signal, nearly two-thirds of consumers begin the “path to purchase” on a smartphone. Among those consumers, 61% continue their shopping experience on a PC or laptop.

What’s driving the use of multiple devices along the path to purchase? The higher conversion rate on tablets and desktop computers can be attributed to convenience, perceived trust and a larger device screen. It thus makes sense that the consumer tendency to conduct research on products via mobile devices would also be based on the convenience factor. Further, the increasing convenience and trust of using apps to make retail purchases and the trend of mobile phones increasingly having larger screens is having a positive impact on mobile conversion rate, although conservatively so.

Implications for Advertisers

These findings beg the question, with the rapid increase in mobile ad spending, what would the most cost-effective ad-format be for mobile shoppers, and would this vary by type of type of item?


In terms of format, mobile video ads have the highest conversion rate of all forms of mobile ads. A study conducted by Trusted Media Brands in January reveals that video ads are expected to replace banner ads as the top mobile ad format by the end of 2016. This study found that increased brand awareness is the primary benefit of video ads on mobile.

It is clear that mobile advertising dollars should be spent in ways that facilitate comparison and research of items. Travel is one of the sectors in which this is evident. In 2016, around half of smartphone users will plan a trip using their phone, but only one quarter will actually book a trip via this medium (eMarketer). Further, given issues with trust and size of screen, it is logical to assume that higher-value items will be researched on smartphones but purchased via a larger device. That said, Criteo research has found that committed app users on average make the largest purchases, as illustrated in the diagram below. We attribute this to a higher consumer trust in apps than mobile browsers.


Based on this study, we believe that cross-device behavior should be the primary focus for both advertisers and e-retailers. For those focused on mCommerce, in app purchases and user-experience in-app should be a strategic imperative. Trust, screen size and convenience in conducting research and comparison shopping are primary criteria driving the behavior of mobile shoppers.

The market has been primed, now is the time for blockchain

by Blake LaFayette on October 7, 2016

VCs have for the most part retreated from investing in bitcoin and blockchain. The appetite for blockchain products however has only increased. Corporations have formed consortiums such as R3 and the Hyperledger Project to learn more about the technology. Numerous enterprises have rolled out internal pilots to test and explore the various applications of blockchain.

At Thomvest, we believe blockchain adoption is about to take off – making it a prime time for VCs to begin making early stage bets.

Take a look at our most recent research report which delves into the current state and future opportunities of blockchain.

An Introduction to Blockchain from Thomvest Ventures

Research Report: The State of App Marketing

by Nima Wedlake on September 26, 2016

Thomvest, in close partnership with our portfolio company Apsalar, recently fielded a quantitative survey to find out more about what marketers are thinking and doing as we finish out 2016 and head into the next year. The survey, conducted in July and August, asked mobile app marketers from around the world about their teams, challenges and opinions on the key issues facing our fast-growing industry. You can download the report in its entirety here, but we wanted to highlight some results around app revenue models and marketing spend.

According to the most recent data from comScore, mobile apps now comprise 58% of connected time in the US.


The figures in many other parts of the world are even higher. Given this, it’s natural that apps are increasingly viewed as a major revenue source for companies. They’ve always been important to mobile –only businesses, but now retailers, travel companies, financial services business and firms in many other verticals are placing greater emphasis on apps as a business channel.

Multiple Revenue Sources Becoming the Norm

One of the first decisions that app makers need to consider is “how will we make money?” By charging for the app? Advertising? In-app purchases? Virtual goods? According to our respondents, advertising has become the most frequently leveraged revenue channel.


Advertising is critical to the so-called “freemium” business model for many gaming apps, where non-payers are monetized through ads. But it is also increasingly common in other categories. Next most popular are real world and virtual goods transactions. Real-world goods sales are the keystone for retail apps, while virtual goods sales are most commonly leveraged in the gaming side of the business. Subscriptions are also becoming increasingly popular, with about ¼ of our respondents reporting that they are part of their revenue models. Subjectively this figure seems rather high to us, which may be driven by our respondent sample. Nevertheless, it seems apparent that subscriptions have “arrived” as an avenue for revenue in our world.

Advertising More Popular Monetization Model in the Developing World

Ours was a truly international respondent sample, and an analysis of developed versus developing world apps showed some interesting patterns. Most vividly as regards in-app advertising. Perhaps not surprisingly, advertising was leveraged more often as a revenue source in developing markets like India, China and Latin America.


Nevertheless, on the gaming side, in-app purchases of virtual goods like game gold represent the most popular type of game monetization.


So that’s some summary information on how apps MAKE money. On the investment side, there is naturally a big range of monthly marketing support levels for apps. About half our marketers said that their average monthly marketing investment per app is $25,000 or less. About a quarter reported spending between 25,000 and 100,000, and another quarter say they average more than $100,000 in monthly spending per app.


To download the complete report in PDF Form, click here.

Blockchain: A Call for Infrastructure, Not More “Killer First Apps”

by Blake LaFayette on September 8, 2016

As our first installment in a series of publications on the blockchain space, we deemed it informative to gather observations around the innovation economy in Silicon Valley, assess opportunities for a new generation of blockchain-enabled systems, and present our thoughts on how the space will develop. To compliment this, below we’re including a deck detailing our initial findings. This reflects a culmination of several weeks of secondary research and conversations with founders, investors, and early adopters. As we continue our research, our theses with regards to timing, infrastructure developments, and use-cases continue to evolve – we hope you find this to be a meaningful summary piece as you explore this rapidly changing space.


It is hard to picture a society without centralized order. Governments, central banks, corporations, and financial institutions all play vital roles in organizing economic activity, enforcing rules, and establishing a sense of trust and order in our lives. With each technological leap forward – mechanization in the industrial revolution, electric-powered Fordism at the turn of the 20th century, and personalized computing and the Internet in recent years – power structures have evolved in response to paradigmatic shifts in.

To illustrate this, let’s deconstruct the contemporary shift to web-based tools. The Internet has enabled a multitude of players to bring our physical world online, and subsequently build platforms for search, P2P marketplaces, conversational interfaces, social media, etc. As a whole, the tech industry has been focused on either improving processes for centralized organizations or enabling distributed interactions via centralized mediums. Hence, the Internet has democratized the open transfer of information, goods, and services. These interactions have increasingly become dominated by large platforms such as Google, Amazon, Facebook, Apple, and Oracle, reinforcing society’s natural affinity for centrally controlled organization.

The valley is known for perpetuating contrarian views which challenge the status quo. The integral role tech companies have played in establishing global organization is therefore counterintuitive. Has centralized authority prevailed because it’s the optimal solution, or rather because it is familiar and straightforward in establishing order? A closer look at the underpinnings of our economy suggests the latter, with centralized services such as escrow, tax filings, bank reconciliation, clearinghouses, accounting, records storage, and countless others relying on painstakingly manual processes. Oftentimes, inefficient processes inflate the cost and latency of transactions while injecting security risks.

In an effort to combat limitations of such internal systems, attempts towards achieving distributed computing have been in the works since the early 80’s, with incumbents such as NASDAQ long demonstrating interest. Blockchain is simply the latest manifestation. Depending upon your level of knowledge and buy-in to recent blockchain hype, you may be an advocate of fully decentralized autonomous organizations (DAOs, made infamous by a recent “hack” of one particular version built on Ethereum), the use of distributed ledgers for certain use-cases such as Bitcoin, securities trading, voting, and land registry, or simply be wondering what the technology is in the first place.

For those of you in the latter camp, a blockchain is simply a data structure that makes it possible to create a digital ledger of transactions and share it among all of the computers in a participating network. Thus, it is sometimes referred to as a peer-to-peer ‘distributed ledger’, and is a tamper proof, public or private, network-hosted record of all consensus verified transactions. Think of Bitcoin as the initial proof of concept for blockchain and posterchild for cryptocurrencies, made famous by regulatory disputes, high-profile hacks, venture investment, and price volatility. To learn more, I’d recommend watching this 6-minute introductory video, reading WSJ’s business-oriented CIO explainer, or looking through CoinDesk’s Q1 2016 report for updated trends and insights.

The recent rise in interest for blockchain is likely due to an oversaturation and consolidation of Bitcoin companies, the narrow market for cryptocurrencies, and the recognition of broader applications of the underlying bitcoin technology, blockchain. In any case, Q1 2016 marked the first quarter in which investment in blockchain startups surpassed that of bitcoin companies, with notable participation from corporate VCs and financial strategics. Moreover, bitcoin companies such as BitFury are pivoting towards offering blockchain solutions at an accelerating pace in order to capture the market’s shift beyond bitcoin.

Consequently as of Q1 2016, seven of the ten highest-funded startups in the space now operate hybrid – i.e. offering a mix of bitcoin and blockchain services – or blockchain-dedicated business models.

Many thought-leaders in the space, including Andreessen Horowitz’s Marc Andreessen and 21’s Balaji Srinivasan, liken the bitcoin blockchain of today to the Internet in the 1990s – a mysterious and undervalued technology that lacked initial mainstream adoption. Yet rather than attributing its slow adoption to a low number of “killer apps” driving usage, our initial research suggests a need for more stable and scalable infrastructure to support the . This refers to those companies building services on top of existing blockchain protocols, most commonly the Bitcoin blockchain or Ethereum smart contracts, in order to deliver solutions for use-cases such as digital authentication, cross-border payments, or securities trading. In a recent piece, famed entrepreneur and investor Joi Ito writes: “I don’t think today’s blockchain is the Internet in 1996 – it’s probably more like the Internet in 1990 or the late 80’s – we haven’t agreed on the IP protocol and there is no Cisco or PSINet.” Lacking adequate focus on improving today’s immature infrastructure, Mr. Ito warns of the possibility that irresponsible deployments will off-put adoption, and that a poor underlying architecture will reduce applications to mere improvements upon narrow use-cases.

Taking a closer look at recent developments in the space, fundraising data offers compelling support for this view. In Q1 2016, the two largest rounds of financing went to infrastructure companies Digital Asset Holdings (abbr. DAH; $62M Series A) and Blockstream ($55M Series A). DAH along with groups making a name by Hence, open-source efforts such as the Linux Foundation’s Hyperledger Project and Ethereum become increasingly significant in enabling collective advances and standardization of protocols. Blockstream offers a particularly compelling solution with its “sidechain” product, intended to enable interoperability between blockchains. As with the Lightning Network’s introduction of payment channels, sidechains will prove critical in improving the scalability of the bitcoin blockchain by compressing transactions to increase throughput.

Two additional developments warrant recognition: (1) product offerings from tech giants, such as IBM, Intel, and Microsoft, and (2) the emergence of three dominant protocols, the Bitcoin Blockchain, Ethereum Smart Contract, and Ripple Consensus Ledger. Many blockchain companies operating within the application layer currently build upon one of these three protocols, with many enterprise blockchain platforms such as Microsoft’s Blockchain as a Service offering cloud-based developer environments for several open-source protocols. Further expanding this assessment of protocols, we encounter an emerging divide between permissioned and permissionless blockchains – colloquially, private vs. public – which subsequently may be either open-source or proprietary.

After speaking with several founders in the space, the consensus appears to point towards open-source as the most effective mechanism by which to encourage widespread adoption and the development of powerful use-cases for blockchain technology. However, many industry consortia groups increasingly operating proprietary solutions built on open-source protocols, such as R3 CEV, are likely to continue strong traction as financial institutions seek to mitigate their risk profile while reducing buy-in costs until the infrastructure becomes more widely proven for its security and scalability.

Given these findings, truly revolutionary deployments of blockchain such as in voting, property title registry, escrow, accounting automation, and digital asset registry are likely at least 12-24 months away, with financial services similarly reducing their risk profile by adopting private protocols in low-risk data areas in the short term. In an effort to consolidate an outlook for the market, the CBDO of DAH noted, “If 2015 was the year we work up to blockchain, 2016 is the year people are running their proof of concepts. 2017 is when you’ll see the first live applications. 2020 is when you’ll see significant adoption, and in 2025 we’ll see a tipping point in some markets. That’s 10 years out.” Our research suggests a similar vision, with a clear emphasis upon developing widely accepted infrastructure in the short-term to help ensure the success of early applications.

Unpacking The Trade Desk’s S-1 Filing

by Nima Wedlake on August 24, 2016

The Trade Desk, a demand-side platform serving primarily agency customers, filed its S-1 on Monday, indicating its intention to go public. Their IPO is an important milestone for the advertising technology sector, which has experienced declines in investor sentiment and deal volume over the last several years. The company is growing and profitable — it reported revenue of $113.8 million and an adjusted EBITDA of $39.2 million in 2015. We took a deeper look into the S-1 and summarized many of the company’s financial metrics in the charts below. For more background on the company, see AdExchanger’s excellent profile from earlier this year.

Operating Metrics

Trade Desk_Revenue & EBITDA

The Trade Desk has experienced rapid growth over the last several quarters. It’s trailing twelve month (TTM) revenue in Q2 2016 was $149 million, a 106% increase over Q2 2015. The company hasn’t sacrificed profitability for growth — its TTM adjusted EBITDA in Q2 was $48 million.

Trade Desk_Gross Media SpendBecause the company’s product offering is primarily self-service, revenue is reported net of all media spend flowing through the platform. It does, however, include gross media spend in its filing; TTM media spend totaled $726 million in Q2 2016.

Looking at the two charts above we can calculate the company’s “take rate”, or the percent of gross billings captured as revenue. The take rate has hovered at around 20% — notably lower than Rocket Fuel (~60%) and Criteo (~40%). These companies clearly have different models, but The Trade Desk is also serving a market segment that is more price sensitive, meaning its take rate could trend lower over the next several quarters.

Trade Desk_Expense BreakdownThe company has done an excellent job managing expenses relative to revenue. Much of this can be attributed to the model it’s chosen to pursue — licensing a self-serve platform to agencies, as opposed to relying on managed services to drive growth. As a comparison, spend on sales & marketing for The Trade Desk represented 24% of revenue in the most recent quarter, versus nearly 60% at Rocket Fuel. We fully expect the company to highlight these differences in its roadshow, along with it’s client retention figure of 95% in both 2014 and 2015.

Trade Desk_Ad Formats

Like most demand-side platforms, The Trade Desk’s ad format mix has transitioned from primarily desktop display to multi-channel & multi-format. The company expects this shift to persist, and plans on investing more heavily in new inventory sources like digital radio, social, native and television.

Comps & Valuation

Trade Desk_Comps

It will be interesting to see how the investor community values The Trade Desk. Clearly, they’ve developed both a scalable and profitable model. But will investors who have soured on advertising technology buy into the company’s long-term prospects? Ultimately, the challenge for investors will be predicting the company’s growth potential given their segment focus. The table below includes some valuation scenarios given a range of revenue growth rates.

Trade Desk_ValuationThe Trade Desk will undoubtedly be a welcome addition to the crop of public adtech companies. The company’s S-1 filing has important implications for a sector that has taken a few punches over the last several years. Their success to date is a testament to The Trade Desk’s ability to execute on its vision and the viability of digital marketing on the open web.

As investors in DataXu, another fast growing DSP serving a different market segment (brand marketers), we’re excited by The Trade Desk’s filing & believe it will pave the way for future adtech IPOs.