Meet ZenClerk, Japan’s automatic revenue optimization platform for e-commerce sites

From the left: DoBoken co-CEO Koichiro Onda, co-CEO Hajime Kuwayama

See the original story in Japanese.

If you recognize the name of ZenClerk, you have a good memory. Born out of the sixth incubation batch of Japanese incubator Open Network Lab in spring of 2013, the team has been developing the service in stealth mode, finally unveiling it last week.

Along with the public unveiling of this new revenue optimization platform for e-commerce sites, its operator Tokyo-based DoBoken announced that it had secured seed funding from four notable Japanese angel investors: Shogo Kawada (co-founder of Japanese internet giant DeNA (TSE:2432)), Masanori Sugiyama (CEO of Japanese social game developer Enish (TSE:3667)), Kosuke Matsumoto (managing director of Enish), and Tomohito Ebine (Chairman of Japanese digital ad agency Opt (TSE:2389)).

According to DoBoken co-founders Koichiro Onda and Hajime Kuwayama, their company has already turned profitable, so the funding aims to leverage the knowledge of these investors rather than filling financial needs.

DoBoken had been initially working on a tool that allows customer representatives to chat with potential customers online, like Singapore-based Zopim which was acquired by Zendesk this year. However, chatting systems require e-commerce owners to allocate some staff specifically for dealing with customers, so the DoBoken team came to believe that these systems cannot always provide user experience suitable to what many potential customers expect. Grounded upon the renewed recognition which shifted their focus to developing a revenue optimization platform for e-commerce sites, they looked to leverage a discount coupon to stimulate buying motivation of online shoppers.


For e-commerce sites, one can analyze behaviors of users from smartphones and PCs by adding one-line Java Script code into one’s website. Based on machine learning and data mining technologies, ZenClerk presents users with a discount coupon at the most appropriate timing and provides the potential customers with the extra encouragement to continue on to the purchase step.

Onda elaborated how the platform is differentiated from similar existing solutions:

Ad networks or ad technology solutions help e-commerce sites catch up with potential customers. But there had been no way to improve the user conversion in the process from that point to the purchasing step. Meanwhile, e-commerce sites had been struggling to find an efficient way to use a discount coupon for better results.

While ad networks and ad technology platforms are focused on presenting coupons that encourage potential customers to buy, we are focused on when to present the coupons to them.

If a check as to how online shoppers behave on one’s e-commerce site could be implemented, one could probably discern when the shoppers have made up their mind about what to buy but may still feel hesitant. ZenClerk refers to that indication ‘Moment’, the exact right timing for presenting users with a coupon so that the sales conversion on e-commerce sites has risen exponentially.

ZenClerk's concept about so-called 'Moment'
ZenClerk’s concept about so-called ‘Moment’

Interestingly, ZenClerk doesn’t present discount coupons to all users. By creating groups of users intentionally receiving a coupon and those receiving no coupon, the platform can conduct an A/B performance test in the two cases. Based on that, ZenClerk can recursively improve the algorithm in detecting the ‘Moment’ timing, and e-commerce site owners can confirm how the sales are improved using ZenClerk via a web-based dashboard interface by comparing the two groups. E-commerce sites will be charged on a performance basis, founded on the sales improvement percentage resulting from optimization platform use.

ZenClerk has been adopted into major e-commerce sites in Japan, such as Dinos. The total amount of monthly revenue from e-commerce sites using the platform has already exceeded 15 billion yen (about $125.2 million).

Collaborative filtering recommendation algorithms have been well known for improving revenue for e-commerce sites. However, it had not always produced results because users were presented with more choices than they expected. ZenClerk enables e-commerce sites to offer promotions automatically, similar to what sales representatives do at real retailers.

Onda explained:

Our solution will not only offer major opportunities in the fashion e-commerce industry but also on any e-commerce sites where users choose items to buy upon preference. For example, the plane ticket may be selected according to price or airline. But regarding where to stay on a trip with your family members or girlfriend, there may be preferences.”KANA? That provides an opportunity for us in terms of helping ticketing sites improve their revenue.

This story brings to mind the Taiwanese startup QSearch which recently won the top prize at Asia Beat 2014, a startup conference held in Taiwan this month. QSearch has succeeded in increasing the conversion performance of Facebook display ads 16 times by targeting ad viewers in utilizing user behaviors rather than user demographics, such as acting with security in mind rather than being in line with age bracket tendencies, for example inclinations to be daring among younger users.

As for a quicker approach to gaining a sales performance, ZenClerk goes a step further than QSearch so there will be huge opportunities for them not only in Japan but also in the rest of the world market.

In this space, we’ve seen Tokyo-based Plaid, a Japanese startup behind real-time analytics tool for e-commerce sites Kartefundraised about $1.5 million from Femto Growth Capital and other investors in July.