2. Session 1 Summary
1. Start with a limited set of product categories, preferably those with
high frequency of purchase
2. Think of how to identify those with an intention to buy
3. Limit the number of options recommended along the select product
4. Identify what are the sources that can be used to come up with top 3
product recs
5. What is the minimum viable product? Consider starting from product
curation
6. How do you get first customers? Why would they come? Who/what
will drive them?
7. Affiliate commissions are tiny – how many users do we need to
break-even?
3. Product categories to start with
● We think of starting from
product categories, which
promise healthy growth rates
while being less exposed to
competition:
1. Apparel & accessories
2. Home furnishings
3. Toys and hobby
● We still need more than one
category - to capture reasonable
range of interests and keep some
variety
4. Identify the intention to buy
Hypothesis: the users who come
from Google search results are
more likely to buy than those who
come from social networking sites
(the two sources seem are top
traffic contributors for many sites)
Why: in more than half of cases
intentional shopping starts from
Google search (check..)
Implications: use Google ads as a
preferred tool to gains conversions Alexa.com
(?)
5. Sources of recommendations
Looks like there are not many Product review site Monthly
really popular product review sites visitors
consumerreports.org 3.4M
In addition to those which
position as “review” sites, we reviews.cnet.com 3.1M
found our that digital media (e.g.
wired.com) and the Amazon itself epinions.com 2.5M
are other very important sources
of product reviews consumersearch.com 1.9M
retrevo.com 940K
We will also seek whether
recommendation sites provide http://techpp.com/2009/03/02/top-10-product-review-websites/
API, so we can use it automatically http://mashable.com/2008/07/18/product-reviews/
(at least ratings)
6. Any idea what’s wrong with Wize.com?
Product curation site
Not many users and
decreasing
Still good downstream
referral rate, Amazon
being top-2 destination
after Google, with 12.4%
share (Alexa)
Alexa.com
7. Minimum viable product
Value proposition: Minimum viable (discuss):
Personalized and curated Personalized and curated
product feeds product feeds. +Limited number
of product categories
Only best-priced offers available Only best-priced offers available
for delivery to your location for delivery to your location
Only the most reputable shops Only the most reputable shops
Sharing your likes and purchases Sharing your likes and purchases
with friends with friends
Premium membership to earn Premium membership to earn
personal revenue personal revenue
8. How do we get first customers
We are thinking of the way to make our customers try and get curious what was
that service that managed to make them a bit happier. That is, first experience, then
attribution. We have discussed several ideas, dropped a couple, and left two for
further discussion:
Seamless Facebook fan acquisition. Facebook already provides customer
filtering possibility through its ad platform. It can be well enough to approach
narrow groups of customers with offerings they are likely to "Like". Seamless: we
create a feed of selected products on our Facebook page, advertise the page with
product image, so if a customer likes a product, s/he becomes our fan (need a
feasibility check). Once we have fans, we’ll have some customer base to work with
and we can target that particular customer sample directly with our value
proposition and drive them to our site
"Social network" for goods. What if goods "liked" their buyers (and probably
other goods, thus creating relationships used for suggestions). So far, we have no
idea how to implement this, but the mere fact of goods liking people might draw
enough attention to our startup
9. Tiny affiliate commissions
We are working on the revenue model now. A very rough estimate
based on 4% revenue sharing (Amazon pays 4-15%) suggests we need to
generate $25mn in online sales to cover a $1mn cash cost with affiliate
commissions. Is it feasible? We think it is, given that US e-commerce
sales amount to almost $200bn
Another thing that gives us some comfort is that the affiliate e-
commerce model is not something new (we list some of the peers in this
presentation)
To address the “break-even point” issue, we also need to understand
our costs, which is even harder than evaluate revenue potential. So
working on the costs too
11. How it works
Folksbay.com
Matching engine
Referral
Product
traffic
Customer preference
aggregation,
processing
curation
Personal information
Friends information
Demographic, social, CJ, SkimLinks, etc.
cultural and economic data
12. Market size
Top-down approach Bottom-up approach:
US 2011 online ad market: $32bn US comparison shopping
Display ad: 35% cumulative monthly visitors (top-5
Performance-based: 65% sites): 60mn
Retail: 22% Assuming:
Intersection: $1.6bn (5.1% of It’s only 1/5 of global referral traffic
online ad market) 1 click a month per visitor
2014: $2.6bn (c.60% 3y growth) 5% conversion rate for shopping sites
Based on 2011 actual data (PwC) and IAB projections 180mn purchases a year
At $20 avg check, it’s $3.6bn
revenue
Indirectly supported by estimates
of the “affiliate marketing”
spending mentioned in the articles
($3-4bn)
Another check: Amazon min
commission of 4% times US e-
See also Market size update at our commerce sales of $195bn =
venture-lab blog: http://venture- $7.8bn
lab.org/blog_posts/1682?team=50
17&data_type=post