Pinterest – Initial Model

Pinterest (PINS.US)

[October 19, 2019]

Initial Model

Download the work-in-progress model in excel here to play around šŸ™‚

It’s been a while since I put together initial analysis and deck skeleton on Pinterest. However, I have some difficulties to convert my thesis (particularly business development and competition) into numeric assumptions and found it’s not easy to validate those assumptions. So I just built an initial model and decided to wait and see.

More detective works required to build up conviction and to determine margin of safety.


 

Pinterest – Initial Analysis

Pinterest (PINS.US)

[August 14, 2019]

Initial Analysis

[0] Analysis Background | Approach Selection

While Pinterest has been a listed company since April this year, it is still in early stage and is more like a pure play of something. If you think runway left is enormous, a bottom up approach to follow the company increases the conviction for a long term bet.

[1.0] How Pinterest Describes Itself and Its Vision

Pinterest is the visual discovery engine to bring everyone the inspiration to create a life they love

[1.1 ] Business Model

Pinterest is an image/video-oriented user generated content site (UGC) with visual discovery engine as key feature for its users to get ideas/inspiration created by other users.

Users generate content by Creating a Pin or by Saving others’ Pins and they can organize their Pins by creating Boards.

create_pin

create_pin2

There’re two interesting features on adding content page 1) Add a destination link; 2) Save from site. Pinterest not only helps users to find ideas/product but also encourage content creators to show where the ideas/products come from. Pinterest help content creators leverage its discovery engine to direct traffic to their sites.

Below is the content of a Pin. Blue circle shows Pinterest’s priority is to encourage to save a pin (engagement and content creation). It also shows which user saved the pin, how it categorizes the item (it’s under “quilted tote bag” board) and how many time it’s been saved. Commercial intent of user can be satisfied as destination link shows this product appeared on Etsy, an e-commerce marketplace. User can follow the link to find the product or similar products available for sale and the shop/person who made the product.

pin1.PNG

While Pinterest has social features that allow users to view other Pinner’s profile, follow them, follow their boards and leverage users’ social graph on Facebook to send message to friends and see their activities on Pinterest… It seems those features were not emphasized when I tried the product on web browser. Apparently, core value of Pinterest is not social networking, news feed based on social graph or communication tool. New user doesn’t need to have a friend on the platform to start experiencing the product. This is important when making comparison between products that have social elements: Facebook, Instagram, Twitter, Pinterest, Whatsapp, Snapchat… Different products have different user behavior which reflects in metric. If someone wants to gauge user engagement by DAU/MAU (number of daily active users divided by number of monthly active users), Pinterest might rank lowest on this metric. This is normal because you would not expect Pinterest’s users to visit the site on a daily basis given current product offering. Higher DAU/MAU might be good and being low might not be a problem.

I feel the core of Pinterest is a crowdsourced visual database where data is either originally generated or saved from internet by users. Users add information or “curate” by adding title and description to each Pin and creating relationship among Pins via organizing Pins under Boards. Those curation help text / visual data science on discovery engine and other features. It’s important for Pinterest to reduce the friction of generating / organizing content, coming up with use cases and finally monetizing traffic and intent on the platform. Maybe Pinterest is a Google in a visual-dedicated user-curated universe.

Theoretically, this platform has direct network effect while it is not as strong as products based on social network.

[2.1]Ā ProductsĀ (Pinterest website) / Features / Function

product1product2product3

[2.2] Ad Formats (Pinterest website) – Monetization

product4product5

[3.1] Traction and Metrics

Pin_MAU

Pin_MAU_Add

Pin_Rev

Pin_QRev

Pin_SRev

Pin_US_arpu1Pin_US_arpu2Pin_G_arpu1Pin_G_arpu2.PNGPin_Int_arpu1Pin_Int_arpu2

[4.1] Geographic Exposure and DemographicĀ 

Around 28% of Pinterest users come from US and the company doesn’t provide further breakdown on international users. Pinterest has penetrated around one third of social networking population in US. The primary Pinterest users demographic in US is female. Currently, US MAU shows saturation. To further penetrate US market, Pinteret might foster categories or come up with use cases appealing to potential male users.

The majority of revenue is generated from US market because its the first market Pinterest choose to monetize. The company also starts monetization initiatives in Canada and some Western Europe markets. Among international markets, I think Japan has huge potential to become a major revenue contributor.

[5.1] Competing Products

Basically, the technical entry barrier is low and Pinterest is competing with large companies that can offer similar features in their product. The moat, if any, comes from Pinterest having what users want and users being able to find what they want or get inspired. Pinterest needs to keep flywheel running fast: User Generated Content – User Curation – Search/Recommendation.

competitor

[6.1] Other Info for Pinterest.com

Pin_websitePin_CompetitorPin_Mkt_ChannelPin_social

[7.1] Reference Model and Benchmark

Pins_SN_userWebsite

Mercari (4385.JP) – Note on Category Deepening, Product Enhancement & Marketplace Stickiness

Mercari (4385.JP)

[December 25, 2018]

* Note on Category Deepening, Product Enhancement & Marketplace Stickiness *

Following series posts on Mercari (initial analysis, withdrawal from UK) and its related topics (network effect and digital payment service), this notes examinesĀ Mercari recent marketplaceĀ deepening/expansion/enhancement initiatives and also does an ideation exercise on user generated content (UGC). The UGC might benefit the marketplace in several ways such as creating more stickiness on user base even though switching cost between platforms is low and increasing traffic internally and externally which might alter purchasing frequency on certain categories and create monetization potential (lead generation, advertising, or even create another marketplace) once the content on particular product/subcategory/category has accumulated to certain level.

[Acquisition of Michael Inc. via Share Exchange]

On October 18, 2018, Mercari announced the acquisition of Michael Inc. which owns automobile-related social network service, “CARTUNE.” CARTUNE has built a community and accumulated automobile related UGC. Mercari aims to increase the listing of automobile category (liquidity) by leveraging CARTUNE’s community. With the benefit of hindsight, the timing and mechanism of the deal benefited the Buyer.

The Company is pushing ahead with its efforts to enhance every category on its C2C marketplace, “Mercari”. The Company, among others, has been making efforts to increase the distribution volume of items in the automotive categories, including the launch of services that allow users to list automotive parts for sale in July 2013 and automobile bodies for sale in May 2016.

Michael has plentiful information concerning automobiles and automotive parts posted by users and off-line and online communities composed of users across a wide range of age groups through the operation of its automobile related SNS service, “CARTUNE”.

The Share Exchange enables the Company, together with Michael, to combine the automotive categories of the Company with the user base, communities and operation know-how of “CARTUNE” of Michael and thereby accelerating the enhancement of the automotive categories with an increase in the number of automotive parts listed for sale.

While the move is quite reasonable and the integration on the technical side should not be difficult (though I’m not a technical guy), the monetization method of marketplace might not be compatible with social network. Simply put, the interaction between usersĀ  fostered by the social network makes it easier for them to transact outside of the marketplace if the benefit of marketplace does not significantly exceed the transaction costs incurred offline, particularly for high value automobiles and automotive parts.

artune

If potential buyers can check onsite and so can they transact offline.

actlist.PNG

It seems Mercari has plenty listings under automobile categories but we can expect liquidity and transactions cluster in certain subcategories. As a market participant, I would be curious about the integration and initial post-acquisition results.

autocat.png

[AI vision proof of concept shopping experience for the Vuzix BladeĀ®]

On December 20, VuzixĀ® Corporation, Mercari, Inc., and Mercari R4D announced the start of a AI vision proof of concept shopping experience for the Vuzix BladeĀ® AR Smart Glasses. (Mercari announcement in Japanese; Vuzix announcement in English)

Below is the prototype demo video.

I don’t see much advantage of wearing glasses to do product search, discovery, or comparison over using the camera on smartphone.

If the stuff got commercialized eventually, my prediction is that it is unlikely to get wide adoption. Mercari will need to spend on customer acquisition and probably subsidize early adopters (price the product below cost or even give away for free). The churn is expected to be so high that the incremental customer lifetime value on power users (heavy users or addictive buyers) cannot cover the upfront customer acquisition costs and the ongoing retention costs (probably in the of form benefits given to users). My use case for this product is customer engagement tool which is free giveaway to power users.

The real asset is AI vision which is able to recognize the item and match with existing listing. Maybe at some point in the future, they will pivot hardware that the software is running on.

[Ideation on UGC]

One of challenges that marketplaces face is that their users can multi-home. In other words, the switching cost between platforms for users is relatively low (while sellersĀ  on the marketplace might find their transaction / review records valuable because it gives them credibility). Unlike technology platform or standard, marketplaces generally do not have lock-in effect on users. This is why marketplaces have to focus on user engagement and mechanism design that makes the platform stickier for users. I think community (social aspect) and UGC are two keys to increasing C2C marketplace stickiness (but still difficult when facing disruptive pricing). As discussed above, there might be conflict between the interaction in community and the monetization method of marketplace. Thus, here I focus on pure UGC where users have to invest their time and efforts to create content, but ignore the social aspect surrounding UGC. The problem of the setup is in reality the ignorance of social aspect might lead to insufficient incentive for content creator to contribute.

While Mercari as a whole has already gained enough liquidity, the density across categories might differ significantly. The UGC experiment should start with specific product that has enough liquidity or users’ interest (via survey).

Below is my illustration of UGC mechanism attached to specific product.

merillustrate

Steps:

0. Decide a category and build an UGC infrastructure that is easy to create, store, present content for creators and easy to consume for consumers. Open the UCG feature to existing marketplace users.

1. Start with 1 or just a few products which either have liquidity or potential buyers/sellers and figure out what kind of suitable UGC types might interest users.

2. Solve the chicken-and-egg problem from supply side by providing in-house created contents first.

3. Observe the reaction, keep records and see whether the feature is gaining traction

4. Expand the UGC feature to other products / categories and open the feature to public.

There will be various curation/modification/fine-tuning in the process. Once the content has accumulated to certain level, the benefits clearly show up.

For example, the UGC on cast-iron cookware are recipe, product reviews, maintenance tips etc. The content can attract outside traffic which might flow to Mercari C2C marketplace and to other websites (Mercari can capture this by setting up a B2C marketplace or incorporate B2C into existing marketpalce). UGC might increase propensity to purchase and the purchasing frequency on Mercari. Consider monetization after UGC has meaningful traction.

Basically, each category, subcategory or even product can be viewed as a vertical. The category-by-category approach in UGC platform building will help the companyĀ  better understandĀ  each category. Probably the company might find better ways to serve some categories. It might be hard for a startup to compete with Mercari on whole C2C marketplace but focusing on particular category is much more feasible. What inevitable is emerging competitors will try to attack on specific vertical and eat away that GMV.

[Side Notes]

When analyzing Mercari as a whole and looking at initiatives it is taking, I cannot help relating it to giant Chinese or US internet companies. Why? Because Mercari is pursuing many opportunities (what they present sounds fancy). This reflects in various subsidiaries and related websites. Coupled with aggressive talent acquisition, I believe the environment within the group must be “vibrant”. While the environment might be chaotic, what it gives Mercari, its shareholders and potential financial/strategic investors is the optionality.

Pursing many challenging opportunities is ambitious but the concept of density / critical mass might be also applicable to human capital / research and development…

~ My Related Posts ~

Mercari (4385.JP) ā€“ Initial Analysis

Networks can still be beaten

Entering a market where existing platforms have network effect working

Mercari (4385.JP) ā€“ Note on Withdrawal from UK

#Mercari

~ Corporate Event ~

FY2019.6 Q2 Financial Results Announcement (15:00 JST)

Google Local Guides

I mentioned UGC the other day and then I got invite from Google Map again to become one of their local guides. I think as long as you post or review something on Google Map, it will try to convert you to become a local guide by sending invite to your gmail account.

I used to checking-in / posting pictures on Facebook and Foursquare but eventually I got churned. Google Map is quite suitable for on-demand / real time site discovery andĀ  location-based search (particular convenient for travelers who have limited local information and that’s why they need guides). The review / rating / uploading information features have been on Google Map for a while but I didn’t notice local guides until recently. Google finds the potential to collect large data by encouraging its uses to generate content and facilitate community building (off-line meetup and online post).

To encourage user to contribute more content to Google Map, Local Guides uses some gamification with non-monetary incentive such as level to reward its content creators.

GMap

Weekly stats report also helps the program keep some user engagement.

WKStats