An E-Commerce CEO

Obtaining funding today is hard. If you’re burning cash, even if you have solid long-term profitability expectations, your business model may have to change.

This week it was Mercado being unable to get financing and then selling for $6.5M to omniture, which was probably a small fraction of the money sunk into the biz.

The lowest risk strategy is to reduce costs without reducing long term prospects. It’s also quite difficult. I’d like to discuss over the next few weeks some methods that may work for your small-mid online retail company.

In this post, I will focus on identifying and reducing your money losing orders.

For every order you can calculate the profitability. To do so you take Gross Revenue (item and shipping charges) and subtract out all the costs associated with the order. The list of costs are long and complex and to manually calculate profitability for a single order takes way too long to be practical. Fortunately it should be a fairly straightforward algorithm which any decent programmer with access to the necessary data can implement.

So, have a programmer implement it (I’ve listed some notes on the algorithm below).

Then look for the outliers, such as orders where you lost money or had your lowest profit margins. Once you’ve identified those orders, try to figure out why they underperformed. I’ve found shipping inefficiencies such as orders being packed in oversized boxes that weren’t supposed to be, pricing errors due to input costs being underestimated and opportunities for greatly improved profitability.

A simplified example opportunity is a high volume item that sold for $37.95 and had a $1 net profit, was raised to $39.95. Volume dropped by almost 20% but with the adjusted net profit amount it’s a money maker.

And that’s the point of this exercise - Maximize your bottom line today.
Your input costs will vary but don’t forget these:

1. Item cost including after discounts
2. Outbound shipping cost
3. Inbound shipping costs
4. Stocking costs
5. Carrying cost - If your average unit of this item sits on your shelf for 6 months, you’re paying storage space and interest on the cost of that item for 6 months.
6. Picking and packing cost for that order - both the labor and the material. Boxes aren’t cheap!
7. Return costs - It’s important to factor this in at a sku level, don’t just take a round number. I have a lot more judo uniforms returned for size issues than I do boxing gloves.
8. Phone costs - If the order is taken by phone that adds significant costs.
9. Manual processing costs - if someone screens your orders, what’s the cost to screen this order
10. Credit card processing costs

Be specific! Generalizing your costs in this calculations will cost you accuracy and keep you from spotting the outliers that we’re trying to identify.

The panel is A Sinfully Engaging Shopping Experience.

If you’re there, say hi.

I have a big conversion rate mover that is repeatable that I can share with you.

Go to Karate Depot.

Clear your KarateDepot.com cookies, then google MMA Gear and click on the link to the KarateDepot.com front page.

Notice the targeted banners and products. (If you do not see the MMA banners, you probably still have karatedepot.com cookies set).

The different front page being served has resulted in a 69% improvement in conversion rate among those who searched for “MMA Gear” and clicked through to our front page.

Excited, we followed up with a test for terms including “Karate” and “Karate Uniforms” and more than doubled our conversion rate for those searches.
There are thousands of terms that lead to our front page but we don’t want to have thousands of customizations for our front page. Our tech isn’t customizable enough to be able to efficiently manage that many different front pages. I can envision a system that would, but for now we’re grouping those terms and creating customized pages for those groups.

Implementation

We set different front page items and banners for different buyer segment in Mercado.

If the entrance keywords match certain rules, we assign the session to the appropriate buyer segment. If the term doesn’t match any of the rules, or if the visitor wasn’t referred by search, we assign the visitor the default buyer segment or add the visitor to an A/B test that we’re running. To determine whether the user came from a search engine we simply look for certain referring urls and then inspect the term the user searched for. For example, a search for “karate” coming from Google will contain “q=karate” in the referring url.
We pass the selected buyer segment to Mercado.

Mercado gives the visitor their customized front page. That front page remains the customized front page throughout the visitors session, to maintain continuity.

Is this for me?

If you have a high volume of visitors to your front page via organic search and some of the keywords are converting poorly, it’s for you.

But my software?!

There’s e-commerce software that has exactly this functionality built-in. We did it with buyer segments and Mercado made it easy. If your software doesn’t have that functionality, ask if they have or can provide a method for creating custom landing pages.

From Here

We’re showing the customer what he wants more often that we were. Therefore we’re selling more product. We will see if we can do the same for visitors to other pages of the site where the page they’re landing on doesn’t have the exact product selection they may be seeking.

We can also take the buyer segment that the customer is assigned for the session and change other elements of the website - such as our related items and others bought sections - to show items the customer might be interested in. Additionally, we can splash MMA branding across the website for our MMA buyer segment and see if we can convince our visitor that for MMA we’re the place to shop for MMA. While of course showing Karate branding across the website for our Karate buyer segment, because for Karate as well, we are of course the place to shop.

Technorati Tags: Ecommerce, Conversion Rate, Mercado, E-commerce

We used LivePerson for a few months, now we’re using SupportTrio.

Why we switched and what we’ve found with both solutions coming soon.