harsh musings (a blog by harsh shah)

I am a huge fan of metrics driven businesses.  Part of it has to do with my love for numbers.  It’s the idea of coming up with conclusions driven by data that is satisfying.

Rental businesses are all the rage these days.  High concept pitches of wanting to be “the Netflix of _(insert consumer good here)_” are popular.  These business are highly metrical and data driven. (Ex: it’s important to understand your customer acquisition cost, the lifetime value of a customer, etc) Chegg, which focuses on textbook rentals, is one of the latest successful models of this type of business.

Two HBS grads recently launched Rent the Runway.  Inspired by an article in the NYT, I decided to take a shot at understanding the unit economics that drive “Rent The Runway.”  Bain Capital Ventures provided seed financing to the business.  The majority of the capital was used to finance inventory purchases (160 styles according to the NYT article).  When your money is being used to go out and buy dresses, you certainly want to understand what kind of cash a dress will bring back to you.  So, I set out to understand the lifetime value (LTV) of a dress.

I’ve made quite a few assumptions in the model. I don’t have access to the site, so I used the NYT article as a base.  The assumptions are as follows:

• Rental prices run $50 to $200 for a 4 night loan.  I assumed that the average rental was the middle point = $125.  I also assumed that the average loan was 4 nights.  $125 does not include the +$5 for damage insurance.  Average rental including insurance is $130.
• In the article, a customer describes getting a $495 dress for the rental price of $50.   I assumed that the average dress costs 10x the average rental price.  Therefore, the average dress costs $1,250.
• 25% COGs margin for shipping and cleaning. This comes out to ~$36.3.  Based on my limited understanding, it costs ~$18 to dry clean a dress.  Let’s say that the Company gets a volume discount for $13.  That leaves $23 for shipping and handling.  I think this is a fair assumption.
• $20 Customer Acquisition Cost per rotation.  I estimated that the Company spends $20 to acquire a customer per rotation.  This is a baseless estimate, but $20 per lead sounds reasonable to me.
• Liquidation value is 10% of the original dress price after two years.  So after two years, a $2,000 dress can be sold for $200.
• A 10% utilization rate, which means that of the possible 90 rotations in a year (360 days / 4 days per rotation), the dress gets 9 rotations per year.
• A 10% discount rate for the cash flow.

Based on these assumptions, the model spits out the NPV of a dress as $118.90.  What does this mean?  For the Company and its investors, spending $1,250.00 on a dress brings you $118.90 in discounted cash flows.

You can find the model here.

At the end of the day, this is a scale and distribution play.  I don’t mean to sound trite, but it is. One of the reasons Netflix was so successful was its ability to establish rev-share deals with the movie studios.  This allowed to the Company to scale its inventory without having to deal with working capital issues that would have otherwise led to its demise.  Rent the Runway could end up struggling with the same issues, unless partnerships are put into place with the fashion houses.  Logistics are even more important.  If someone doesn’t get their bridesmaid dress in time, all hell could break loose (as Jeff Roster from Gartner accurately points out in the NYT article).  This is unlike a late DVD or textbook.

You should view the model as a thought exercise.  It’s likely pretty inaccurate, but I think I’m including the appropriate considerations in thinking about the business.

The model is quick and dirty. Please don’t judge me - I’m a recovering finance guy. I’m doing this while on vacation.

I would love to hear your thoughts and comments.

After a great year and four months at Hercules, I’m excited to say that I am joining the Product Marketing team at AdMob.  At AdMob, I’ll focus on the development of the AdMob Metrics Report.  The Metrics Report is a monthly look inside the traffic flowing through AdMob’s network.  The Report has become a big piece of AdMob’s brand and is used widely across the internet as a barometer to measure the growth of next generation mobile platforms (ex: GigaOm, Mary Meeker’s Web 2.0 deck, Ad Age).  As part of the position, I will also indirectly support sales and biz dev activities.

I’m thrilled to start in an operational role in a space that I’m passionate about (mobile), working with a team that I’m excited about in an area that I think has yet to hit its inflection point (mobile advertising).   It’s also exciting to be returning to marketing, a field that I touched in undergrad and at Mahalo.  At UCLA, I ran the Research Sub-team for the UCLA Advertising and Marketing Team.  In addition, my BA is in Communication Studies.

One week ago Google entered into an agreement to acquire AdMob.  With Google’s backing, AdMob has greater tailwind to “accelerate the pace of innovation” as Omar H. put it, and go out and grab more market share.

I’m ecstatic at the opportunity to be on the front lines of the development of the mobile internet.

I want to thank Mike Fyall, Jason Spero and the rest of the team at AdMob for the opportunity.

This is my favorite article of 2009 (thus far) from the New Yorker.

If you believe that capital punishment is an effective form of disciplinary action, I implore you to reconsider.  The risk of executing an innocent person is just too high.  More importantly, it’s not humane.

There’s a lovely article in the NYT about mistakes in typography disturbing typography purists.  Details are important to someone.  Even the smallest details.  Someone is out there watching and judging.

I am very excited about the launch of Typekit.  I haven’t seen much innovation around typography and Typekit’s model is particularly interesting.  If Typekit can monetize typography on the internet, imagine the the total addressable market for typography on physical goods including books.  +Advertising copy.