How to Finally Nail Philippine Online Retail
Or Cebu Startup Weekend Pitch Take # 2
May 2012
Back in the 90’s, the biggest buzz word was eCommerce. The US had the likes of Amazon and Webvan. The Philippines had MyAyala. Amazon suceeded and Webvan made a crater so huge, it will be enshrined in countless MBA case studies in years to come.
In the Philippines only one eCommerce model has thrived. You sell to overseas Filipinos and deliver locally. MyAyala has become MyRegalo, following this model.
Every chance I get I ask people in the business why eCommerce isn’t working here. I usually hear two things:
- Low credit card penetration (the last figure I saw was 5%)
- Filipinos need to touch what they buy
So the question is still open. Will classical eCommerce work in the Philippines? (ie, eCommerce website + online marketing + distribution center + outsourced delivery). Have market conditions changed since MyAyala?
Thanks to the Samwer brothers we will find out in a few months. Earlier this year, Rocket Internet brought Zappos and Amazon to the Philippines (they call it Zalora and Lazada). The Samwers are known to cut clean if a business doesn’t fly in six months.
India’s Flipkart seems to be doing okay, so the Zalora bet isn’t unreasonable. Still, a man cannot help but ask: is there any other way to play this game? Is there a way to take credit cards out of the equation and to let buyers touch before they buy?
Playing a different game
Imagine this. A girl orders a dress online. Instead of outsourcing the delivery, we deliver with a mobile store. Maybe she buys what she orders or maybe she doesn’t. It doesn’t matter. The important thing is that we put inventory in front of someone who has the probability of buying.
Furthermore, we can up-sell at the point of delivery (vs Amazon and Zalora who could only up-sell in the web). No pre-payments needed. We balance the risk of incompletes with the opportunity to up-sell.
So the game now is not about efficiently fulfilling online orders, but bringing inventory to where it has the greatest probability of being sold.
These mobile stores already exist, but they are not connected to an eCommerce website and are not fueled by online marketing. As part of business validation in the Startup Weekend, we showed images of these to 21 women. I’ve been doing customer validation for more than a year for another business and I see a certain twinkle in the eyes of my customers when I mention a jackpot idea. The eyes of women light up in the same way when we show pictures of these vans and explain the concept. They love it.
Proximity-selling
To maximize sales from your delivery cost, you want to sell as much as possible in whatever area you are in. Logistics companies plan out deliveries to follow an efficient path. You can also do this in our case. And since you are bringing your inventory with you, you can also do some technology-enabled impulse buying.
The easiest way is to announce via Twitter or FB your delivery route. People reply that they want to be visited. You go there and increase your sales. You could also build a geolocation app which alerts when our mobile store is nearby. With one click, you visit the app user (whose address is already in your database). See uber.com for a technology precedent (thanks Taylor Miles for this tip!).
In the Philippines, to forget SMS is to throw away opportunity. We brought in some traffic to a “download” page for the proximity-alert app. The results show that the Philippines is still SMS country. (Here’s a startup idea: work with Telcos to create a location-based opt-in alert system using SMS. Telcos can geotarget mobile phones using cell site location. Why not monetize this ability? I’m paying PhP2 — PhP30 for PPC. Businesses are probably willing to pay a premium for SMS targeted to customers in relevant locations.)
Back of the envelope computations
The rent for a clothing store is ~PhP 50k or ~15% of revenues. If you depreciate a van worth 1M across 3 years, you pay ~PhP 1K a day for it. Just add maybe PhP500 a day for gas, and your cost would be roughly the same as a boutique in the mall. Your cost for personnel, business taxes, etc should roughly be the same.
Here’s another way. According to one clothing line owner, they give a cut of 30% to boutique owners. Their clothes cost around PhP1,500. If you just sell 3 clothes a day, you already pay for your van and fuel.
The big question is whether you could more clothes per day than a static boutique. No revenue projection survives first contact with customers. So the only way is to prototype the business.
Malls discriminate against call center agents
Aside from overseas Filipinos, the heroes of the Philippine economy are call center agents. So it’s strange how malls discriminate against them. Malls close when they are awake and open when they are asleep.
On the other hand, our mobile store could sell 24x7. On weekends, it could go near residences. On work days, it could go to the offices of working women. On work nights, it could go to offices of women who live in the US time zone.
Delivering to offices should be taken seriously. In the website that we used to verify this idea, we asked for preferred delivery location. Surprisingly, only half replied “home.”
Slay Goliath with Moore’s law
Planning out the route of your mobile stores in order to maximize sales will be an interesting challenge. Lining up the orders efficiently is just the beginning. Which van has the right inventory for specific orders? Among the orders and the people nearby you will alert, who has the best probability of buying the most? Which locations tend to buy more and from which stores? How do you optimize route considering traffic, weather conditions, day of the week? How do you optimize inventory based on sales history, online “Likes”, “your friends also bought,” etc?
You need an olympic-level data scientist to develop this baby. But once you do, you have Moore’s Law on your side. While real estate (ie, rent in malls) becomes more and more expensive over time, the cost of your route and inventory optimization system stays flat as you scale and becomes cheaper over time. You could slay SM Megamall with Moore’s Law. (See Bill Gross using Moore’s law in the solar panel business).
Once you have a network of mobile stores, you are actually acting like a mall which is not limited by space and which could be optimized using historical sales data. In buzz-word lingo, this business is: A spatially-flexible mall dynamically optimized through eCommerce, proximity-selling, and historical sales data.
Proof of concept
In a Startup Weekend, you have to verify your business idea. We interviewed lots of women and observed online behavior for a “concept website.” Aside from the graphs above, here are other relevant data:
- Traffic cost: PhP 7,495.1 (Thanks Multiply and Google Adwords!)
- # Visits: 1,166
We sent that traffic to two websites
Results for http://cebu.oliviaandfifthph.com/
- zero orders
- 14 sign-ups to be visited by the fashion store
- 2.37% conversion rate
- PhP178/sign-up
Results for http://themobilefashionstore.weebly.com/
- 33 sign-ups to be visited by the mobile store
- 5.56% conversion rate
- PhP161.10/sign-up
When you reach the thank you page of website #2, you are given an additional form which offers a “download” of the proximity alert app above. Out of the 31 who reached that page, 25 downloaded (that’s an 80.64% conversion rate!)
To simulate the proximity alert app, we texted the 35 women who gave their mobile numbers. 5 replied (14.39% reply rate).
- 1 wanted to be removed from the alerts
- 2 wanted to be visited but some other time
- 1 wanted to be visited immediately
- 1 wanted to buy a mobile store
I’ve been doing conversion-driven online marketing in the past few years. These results look pretty good to me (except perhaps the cost per conversion — but this could easily be halved after doing some PPC cleanup).
We were not able to test a few very important things: eg, the success rate of deliveries using a mobile store, the average order size, the average up-sell size, and the average # deliveries per day. These are the metrics that will make or break this business. And these are the metrics that could be tested only through prototyping the business.
The 3rd Retail Revolution
This was the name of our team in the Startup Weekend, although we never got a chance to explain it. So here it is.
Megatrend 1: Massive urbanization in the late 20th century
Retail Revolution 1: Super mega malls
Megatrend 2: Invention of the internet and credit cards
Retail Revolution 2: Classical eCommerce
Megatrend 3: Ubiquity of mobile devices and location-based apps
Presenting… The 3rd Retail Revolution!
Business models are public domain. Only execution is proprietary.
I’ve had a minor obsession with this idea for around a year now (I could not prove to myself that this won’t work!). Although I’m prioritizing my ongoing business (with paying customers), I’m really interested in seeing whether this will actually work. I invite you, dear reader, to pursue this. Who knows, you might be the next retail tycoon (when you become such, just give me a dozen or so mobile stores for giving you the idea :-)
***
Special thanks to my #swcebu team mates: Jake Morales, Peter Corazo and Hannah Panares. Thanks as well to Gem Co and Denise Young of Olivia & Fifth, and to Rob Barretto. Thanks to Danna Lee, Kyne Santos and Hannah Tan for the Adwords setup and optimization. Thanks to Multiply.com and Google Adwords for the ad credits. And thanks to the organizers, mentors and judges of the 1st SWCebu! You are our heroes!
Originally published at www.corazo.org.
