Serious Retail Technology and More

A picture is worth a thousand words.... and a few dollars!

Nov 21 2008 3:35PM

As I mentioned in my last post, in my spare time I have built (and operate) an online business known as beautycrunch. You can read details about the business at beautycrunch.com, or in the previous post, but it's basically an authorized direct to consumer reseller that offers overrun and discontinued (but still very good) artistry brand beauty products at a great discount.

It's a brisk, fun discount business, but there are some financial challenges since the margins aren't near as great when selling product at full retail. Understandably, all operational costs come out of my pocket, not the brands. I pay for all the packaging, shipping charges, credit card processing fees, charge backs, mistakes, etc,… The brand get a greater % take, so for me to make money I have to do everything in my power to make sure every order goes out perfectly - I can't afford to stay in business if I'm unnecessarily reshipping orders.

What do you mean unnecessarily reshipping orders?

My average item count on an order is better than most retailers, so what do I do when a customer complains that they received their order but an item is missing? Sure, I make mistakes, but you might be surprised at how many times I don't!

So why would someone say they didn't get an item when in fact they did?

1. Roommates steal stuff.
2. Small items can get lost in the packing material.
3. Some folks are just not as honest as you and I.

I like to think I have very few of the number “3” customers, but no business can completely avoid them.

This may sound like a trivial problem, but it's not. It's a big loss for any retailer when this happens, and can really slow down and even stop the growth of a small startup. There are the obvious cost for the additional product, not to mention the extra shipping charges and shipping supplies. And of course there is all the wasted time! And if you get a reputation for a company that just reships on customer complaints alone, you might be surprised to see how many more you get!

Currently I don't have a need for a sophisticated WMS package that prompts me to scan UPC labels before shipping, but I do have a digital camera and a $7 whiteboard. This is not a technically impressive solution, but I've saved a significant amount of time and money by simply taking a digital picture of every order before I pack it out.

Once a week I empty my camera's memory card in a new date labeled directory.

Tech tip: date label your directory like YYYYMMDD and you'll always be able to sort it!

If a customer complains that an item is missing, I simply email them the photo of what I packed out. And truthfully, the extra step of taking a photo prompts me to double check the items picked. Admittedly, there have been times where the customer is right, so it's a win for the good customers too!

Below is how this usually plays out these days. I skipped the first email, since it was just a message reporting that an item was missing in their order.

Subject: RE: Missing item
Date: Fri, 15 Aug 2008 13:09:38 -0700
From: info@beautycrunch.com
To: << email removed >>

Hi << name removed >>,

Thank you for your email.  We researched your order and it did ship complete 
(see attached photo).   We photograph every order before it leaves our warehouse.   
Please check under the bubble wrap to make sure it isn't just buried in there.  
If not, please check with any neighbors, roommates, etc. who could have intercepted 
the package before you received it. 

Thank you,
The BeautyCrunchers  

-----Original Message-----
From: << email removed >> 
Sent: Sunday, August 17, 2008 9:45 PM
To: beautycrunch
Subject: RE: Missing item

Thank you for the picture, it saved my life.  Turns out my roommate/sister had taken 
look #1 while I was putting the scissors away, she denied having it but after showing 
her the picture of the shipment and telling her that no one else could've taken it she 
gave it back.  Thankfully she hadn't used it yet.  Anyway, sorry for blaming your 
company on not having sent it.  I love your site and can't wait for new items to 
come in.

<< name removed >>

Now did I really save her life? Maybe.

Again, this isn't the most technically sophisticated tip I'm sharing, but it is one that could save the smaller retailer some serious $ and assist in improving their pick/pack process.

Agree? Disagree? Post a comment, or email me - maybe I can add some updates to this topic if I skipped something super important.

How can you not take credit cards at a trade show?

Mar 6 2008 5:12PM

In my spare time I have built (and operate) an online business known as beautycrunch. For background purposes, beautycrunch is an authorized direct to consumer reseller that offers overrun and discontinued (but still very good) artistry brand beauty products. beautycrunch carries premium brands like Stila, Too Faced, DuWop, and Lola that attract artistry enthusiasts who are looking for a “better than good” deal.

Most of the time I'm selling direct to the consumer online, but every so often I'll participate and sell product at a conference, trade show or special event. I needed to accept credit cards and have a speedy checkout process, but also wanted to keep the technical cost structure low. For me, there where were only two components required for selling offline at remote locations, and I found solutions for each worth sharing.

A. reliable, low-cost Internet access

B. low-cost, simple point of sale (POS) application that supports a USB magnetic strip credit card reader, prints a branded customer receipt and interfaced with my online payment service (PayFlow Pro/PayPal).

Some events simply don't offer Internet services, and the ones that do provide those services at insane rates, often over $1K for a small block of days (and poor shared bandwidth). And getting any connectivity support during an event is hopeless.

Since I only need this type of Internet access a few times a year, locking myself into a 2 year Sprint/Verizon type mobile wireless broadband card subscription wouldn't make economic sense. So, I simply rent an already activated wireless card from RovAir. I can only say positive things about this company and the service they provide. I've used their service a few separate times now and they have never missed a date... and the card worked perfectly.

It cost just under $60 for a 3-day rental. And the wireless card return works like NetFlix; RovAir includes a preprinted return UPS envelope in the shipment. Very simple.

For a low-cost, simple point of sale system, I recently started using USBSwiper. It's good for remote events where you need clean printed invoices, SKU tracking and fast (swipe-able) credit card transactions that seamlessly interface with your PayPal Payments Pro or PayFlow Pro account.

Keep in mind that it's not a full featured POS - don't expect stock transfers, fancy gift receipts, split tenders, fancy reporting, etc…

It's created with FileMaker Pro, so the interface is a little clumsy, and certain functions are not near as intuitive as they should be. I have a large list of items (wishes) that would make this product significantly better for my use (auto print invoices after sale, better messaging when transaction fails, some reconciling and simple reporting features, etc,…). However, conceptually, I think this product has real potential.

For around $50 you can buy a USB magnetic strip credit card reader almost anywhere, including USBSwiper. Just make sure it's a Track 1 and 2 reader with keyboard emulation.

I believe USBSwiper is a one-man operation (Andrew), but he is very responsive and seems committed to making his application better and better. I personally think the next move would be to move away from FileMaker Pro... but, hey, it works well enough and the price is perfect at under $5 a month. Nice!

So for a total of about $65 for a 3-day event, I am the most technically sophisticated vendor providing the fastest and coolest checkout!

Agree? Disagree? Post a comment, or email me - maybe I can add some updates to this topic if I skipped something super important.

Pedestrian Counters

Mar 4 2008 11:43AM

Just as you measure traffic and conversion with your online store, the same can be done for your traditional brick and mortar stores. There are just a few players in the pedestrian counter space.

ShopperTrak [link]
St Michael Strategies (SMS - formerly know as BMI) [link]
Traf-Sys People Counting Systems [link]
Counting Technologies Corporation [link]

ShopperTrak offers the “Orbit” video camera setup and makes claim to distinguish between incoming and outgoing shopper traffic, count side-by-side traffic and differentiate between adults and most small children. Nice, but I don't really know how important all that stuff is. To me it's a bit like the toaster that can also cook you an egg. Great, but I really only need some toast.

When it comes to foot traffic, the only data most people have time to care about is door traffic and conversion. They'll just quickly look at those numbers and compare it to like doors and last year's numbers. Every counter solution will tell you traffic by hour, so figuring out your prime and weak selling hours is easy.

What is important is that even if it's not 100% perfect, as long as you're calibrating to yourself it doesn't matter all that much. What retailers look for are changes in traffic patterns and dips & surges in conversion.

ShopperTrak is the more expensive solution (setup, monthly fee and annual maintenance). The Orbit units requires special installation and not too many people can fix them when the break. However, ShopperTrak offers some cool things like their National and ShopperTrak Retail Traffic Index data, which benchmarks and projects retail customer traffic trends across different segments. That predictive and historical information could be quite helpful.

Traf-Sys offers the budget conscious Door Mounted Infrared Sensors, as well as an overhead body heat monitoring Thermal Sensors. They also offer a Proximity Sensor, which will tell you how many people are coming into a particular area of your store.

Your traditional infrared style door mounted traffic units count intersections in .5 increments; if you walk in the store it's .5, then when you leave it's another .5. So, .5 + .5 = 1 whole person. Some retailers will count it to be slightly less than 1 since deliveries and employees often use the same doors.

If you have the option, for door mounts I would recommend you go with the low-voltage wiring as opposed to the battery option. The enticing battery option is an easier setup, but batteries die and having store associates replace the batteries and try to re-sync them can be nightmarish.

Mount the units at lest 4' high to minimize the stroller and small children counts. Unless you're a candy or balloon store…then maybe you do want to count the kids.

I recently read an interesting post on TechCrunch about UK based company, Path Intellegence. Path Intellegence's FootPath™ solution seems like a clever, promising new way to monitor mall and store pedestrian (foot) traffic. Cell phones periodically ping nearby cell towers with a unique ID — Path Intellegence has developed a monitoring technology to pickup that ping and continuously (and accurately) track the anonymous movements of that shopper throughout a large location.

Path Intellegence is unlike any other tracking device in that it promises to also track flow of traffic (hot spots, bottlenecks, etc,…), which would be extremely valuable information for any store based retailer.

Maybe someday malls will adopt this solution and provide Path Intellegence as a service to their tenants on a subscription basis. Tenants would then not have to install or maintain their own tracking devices and could even get mall traffic to correlate with their individual store traffic data.

Agree? Disagree? Post a comment, or email me - maybe I can add some updates to this topic if I skipped something super important.

URL Rewriting

Feb 28 2008 8:44PM

From what I have heard and read from today's top Search Engine Optimization Experts (bahahahaha!), the semi-intense maintenance task of URL Rewriting (a.k.a. friendly or clean URLs) could be worth the effort. The word on this old subject is that search engines put some indexing weight to descriptive relevancy or keywords included within a URL. And some search engines altogether avoid complex URL structures that could result in loops and site overhead.

There are some security benefits to URL rewriting, and they look much cleaner, but I think the most compelling reason to create friendly URLs is that they're more useful to email and/or save if they have some obvious context to them. And it's pretty easy to do.

For a retailer, it makes sense to at least have your department and category URLs search engine friendly and clean. Departments and Categories don't change all that frequently, and most retailers don't have an unmanageable amount of them…. and category and department names are often popular search terms.

Looking at GAP's baby infant outfit category URL...

You can see that baby infant outfit category is 35358 (cid=35358), but a friendlier URL might be...

http://www.gap.com/browse/category/baby_outfits

(that will be $300 for my SEO consulting service, GAP)

I see that the top natural search positions from Google usually have some URL search content within the link, so it could be worth the effort of creating and maintaining the URL mapping. But why would GAP (and all their other brands) not be doing it? Have they not chatted with an SEO expert?

This simple 4 page site (created in C# .NET on the FREE SQL Server Express 2005) uses URL Rewriting. If you navigate to the previous post, you'll see the URL as…

http://retailbyte.com/how_to_minimize_online_fraud_and_abuse_of_marketing_promotions.aspx

However, the real link is…

http://retailbyte.com/?ThreadID=10

ASP.NET 2.0 features a limited, static URL Mapping tool. There are a lot of tips and downloads for creating extended HttpModules with Regex support, but your static category and department mapping project might not be too hampered without that functionality.

I downloaded UrlRewriter.NET and had everything running in just a few minutes. Read Scott Guthrie's blog post to get a good technical explanation of some of your URL rewrite opitions here.

If you're running on Apache, the mod_rewrite module is pretty sophisticated and seems to be what developers running IIS are chasing.

Agree? Disagree? Post a comment, or email me - maybe I can add some updates to this topic if I skipped something super important.

How to minimize online fraud and abuse of marketing promotions

Feb 22 2008 5:44PM

In an earlier post I wrote about business rules for processing online orders. Tight business rules, with well thought out supporting software, are the most important component in significantly reducing fraud and promotional misuse. Another important tool is a smart Order Management application that exposes all aspects of your orders and shoppers.

(my next post will be the perfect Order Management Application… which you can have for FREE)

Business Rules: There are many common business rules that can be applied across any retail business, but many are unique to your particular business. For those (sadly) most discover them the hard way.

Below are some easy, cross-market rules you should consider…

1. Use an Address Verification service along with your credit card authorization. I authorize offline, but flag for review any order that doesn't come back with an address and zip match. I am a big fan of PayPal's PayFlow Pro product (formally VeriSign's).

2. If you only ship domestically, make sure you flag orders coming in from IP addresses outside the U.S. That doesn't mean you cancel the order, rather put it in a bucket so your customer agents can review it before releasing to your distribution center.

For about $600 you can buy a comprehensive IP database from IP2Location. It's a simple one table import that allows you to pull details down to the latitude/longitude of the ISP. You can also use this same table data to serve up regional content at your site (example; store events in your area). Without question, it'll pay for itself almost immediately.

3. Large dollar review is a master of the obvious business rule, but every business differs on what constitutes a large dollar order. A good approach is to define different dollar thresholds based on the customer type. For example, repeat customers that have shopped before (without issue) should have a larger dollar threshold than your first-time customer.

4. Look at new customer orders that are shipping priority (Next Day and 2 DAY).

5. Look for a burst of same-shopper activity. If a single shopper shops multiple times in a short period of time, flag ALL those orders for review.

6. Look for a burst of activity from a single IP.

7. You should flag for review orders that indicate atypical shopper behavior. Over time you'll know what atypical shopping behavior is for your business — take some time and define that behavior and implement those rules.

A few examples…

If you have an obvious active promotion (coupon) and a sizeable order didn't leverage it, be sure to flag that order for review. A real shopper is usually going to take the time to enter a coupon to get the best deal possible.

I've seen quite a few fraud orders come in with obvious bad phone numbers. It's worth reviewing orders that have area codes of 000 or 123.

etc,…..

8. Create a “never ship” table in your database and compare your new order addresses (both bill to and ship to) against that list. If you get a chargeback or experience fraud activity from an address, in a hash like form, store that address and order in a simple indexed table and compare every new address (via your hash function) against that table.

If you have a ONE per household type offer, this is a good way to catch the violators. Most retailers associate these promotions to the shopper (email and/or unique shopper_id). Sneaky shoppers will create multiple profiles to get over on the ONE per household offers.

I take the first 5 ZIP and the ADDRESS, then remove all non consonants and minor address descriptors (RD, ST, AVE, STREET, CIR, etc,…)

1234 E WOODLYN AVE
PASADENA, CA 91104

Becomes…

911041234WDLYN

TIP: you can also use this hash to find out (roughly) how many duplicate shoppers you have in your database

Here is a simple VB sample function of how you might go about creating an address like hash. You can make it much more powerful (in almost any language) using regular expressions, or even do this at the database level.


myAddressGUID = CreateAddressHash(RemoveDescriptors(strZIP & strADRESS1))

Function CreateAddressHash(ByVal strInput As String) As String
Dim x As Integer
Dim strResults As String

    On Error GoTo Err_Handler
    
    For x = 1 To Len(strInput)
        strResults = strResults & IIf(InStr("1234567890BCDFGHJKLMNPQRSTVWXYZ", _
                                  Mid(UCase(strInput), x, 1)) > 0, _
                                  Mid(UCase(strInput), x, 1), "")
    Next
    CreateAddressHash = strResults
    Exit Function
Err_Handler:
    CreateAddressHash = Err.Number & "|" & Err.Description
    'log error
End Function

Function RemoveDescriptors(ByVal strInput As String) As String
Dim strReults As String
    strReults = UCase(strInput)
    strReults = Replace(strReults, " RD", "")
    strReults = Replace(strReults, " STR", "")
    strReults = Replace(strReults, " ST.", "")
    strReults = Replace(strReults, " STREET", "")
    strReults = Replace(strReults, " ST", "")
    strReults = Replace(strReults, " CIRCLE", "")
    strReults = Replace(strReults, " CIR", "")
    strReults = Replace(strReults, " AVE.", "")
    strReults = Replace(strReults, " AVE", "")
    strReults = Replace(strReults, " BLVD.", "")
    strReults = Replace(strReults, " BLVD", "")
    RemoveDescriptors = strReults
End Function

Once your rules are defined and implemented in your order flow (authorization) process, you need a smart Order Management application that allows your team to quickly manage and review orders, ultimately making the right decision on how to dispense an order.

My next post, The Perfect Order Manager, will cover this in greater detail.

Agree? Disagree? Post a comment, or email me — maybe I can add some updates to this topic if I skipped something super important.

Google is one very costly spell checker

Feb 11 2008 7:02PM

I'm not a great speler (sic) and I don't like typing full URLs in my browser address bar, so like many others I've come to use Google as a spell-checking bookmark.

I rarely use browser bookmarks for sites I don't frequent daily, especially when online shopping. For those sites it's easier to type some word jumble into my Google toolbar and have them correct it for me. They'll even give me a clickable link. Nice! But if you're clicking the top (sponsored) link, someone is paying for it… and there could be misleading reasons why they think they're getting their money's worth.

GOOD GUY DISCLAIMER: I'm a thoughtful guy and I do my best to avoid clicking Sponsored Links if the natural results link is on the same page.

Here is an example:

I'm looking to buy some shoes at amazon's endless.com. You can see that I incorrectly spell "endless" - Google corrects it for me, then gives me a sponsored link that'll take me right there (thanks Sergey).

When I click through, Google's cash register rings and endless.com pulls the referring site ID from the query string in the URL. Master of the obvious stuff, but retailers do this so they can pass that to their analytics software, as well as store it with their orders to measure keyword effectiveness and validate conversion and such. It's an ROI thing.

WTF: Why is endless.com still in BETA? Haven't they been around for a year or so now? And why would they need a BETA in the first place? They sell shoes, right?

A minor point, but if it's truly a first-time visitor, and your site is what they were hoping for, there's a chance the new visitor will create a bookmark as soon as they land (see my endless.com bookmark below). Now every time the user selects this bookmark the retailer could see an inflation of Google's value… unless the retailer is savvy enough to be retiring keywords and creating strong affiliate reporting filters.

You can mitigate that effect by doing a quick self redirect when the user hits your landing page with an affiliate parameter in the URL, initially tracking the referring tag but removing it on the redirect before the user has a chance to bookmark it.

Be smart and study your online repeat customer behavior. If your repeat customers keep coming your way via costly Google keywords (ones that are obvious to your brand), chances are... they too are using Google as a spell checking bookmark.

Agree? Disagree? Post a comment, or email me — maybe I can add some updates to this topic if I skipped something super important.

Secrets of Online Conversion

Feb 9 2008 10:48AM

Trying to find reliable info on e-commerce conversion rates is a notoriously frustrating quest. Most companies are highly protective of this information - and they should be. Apart from your balance sheet, there's probably no greater indicator of your business health than this metric. Thus, very few online retailers will open their kimonos wide enough to reveal the goods. Thankfully, there are a few exhibitionists out there, and I've seen enough on my own that I'm able to help shine some light into this dark mystery.

First thing to remember about conversion rates is that everybody lies. Okay, maybe they don’t lie, but they’re definitely going to put their best spin on it. Whenever someone tells me their conversion rate, I silently divide it by two. It’s not that there’s a conspiracy to spread misinformation, it’s just that most of the people with access to conversion rates are the ones whose jobs hinge on them. Thus, it’s safe to assume the conversion rates you hear are inflated.

So how do people inflate their conversion rates? It’s all in the methodology. I believe the most honest way to report conversion would be VISITORS to BUYERS. And depending on what vertical your business is in - these can vary dramatically. For example, a fashion retail website should expect average conversion in the 2 to 3% range. Someone selling high-ticket electronics probably sees much lower conversion in the 1 to 2% range.

But remember, everybody slants their numbers. Some e-commerce companies with content-rich websites like to segment out traffic that loiters on their non-product pages and they don’t factor those into their overall conversion formulas. Some companies go even further, and only factor conversion based on those shoppers who actually add something to their cart (thus they are officially "shoppers") against those who ultimately make a purchase. Conversion based on the SHOPPER to BUYER method is a good gauge of how your site funnels people through the checkout process, but it doesn’t tell you much about the unrealized potential of your overall visitors.

Obviously certain segments of your customer base will convert at higher and lower rates. If we just carve out the repeat shoppers (people who've made more than one purchase) then conversion should probably be around 2 or 3 times higher than the overall percentage. This shows that the shopper is happy with their previous purchase and the barrier to convincing them to go through with an order has been lowered because you delivered on his or her expectations. A very scary thing to see is when your repeat conversion is flat to, or lower than overall conversion. This indicates dissatisfaction with the product they received (maybe a quality control problem), or perhaps a serious usability issue on the site.

What are the Top 3 culprits behind crappy conversion? Inventory, inventory and inventory. To prove this, keep track of your in-stock ratio and plot that against your conversion rate. You will find that even a slight drop in the availability of your best-selling items will always correspond with a lowering of overall conversion. It’s the old Pareto Principle (A.K.A 80/20 rule), where 80% of your sales come from 20% of your products. Remove one of those items from the equation and you’re hurtin’ for certain.

Timing is another factor that can produce a fluctuation in conversion rates. You’ll see this clearly demonstrated in email campaigns. What time of day is the email landing in the recipient's inbox? One rule of thumb is to avoid the Monday morning clutter zone. This is that time starting at 9AM on the east coast when everyone returns from the weekend and they begin a fast and furious frenzy of deleting all their accumulated spam. You don't want your email to be a casualty in this kill zone, so you're better off landing your campaign later in the morning. This tip alone can boost your open rates dramatically.

Without a doubt, the best way to move the needle on improving conversion is to take a very close look at your checkout chain. Starting at the cart and moving all the way to your “Thank You” page, find ways to streamline this process down to the bare minimum of clicks. Promotional offers are easy triggers, but there are obvious brand implications to that practice (another topic for another day).

Now that I’ve muddied the water with all my disclaimers about how everybody cooks their books on conversion, I’ll give you what you’re probably looking for…benchmarks! The nice people at Fireclick have provided one of the only public places I’ve found that actually attempts to track conversion across different online verticals. But be warned, you should condition the data you find here with all the information that I’ve shared above:

fireclick.com

VIP Cards - very smart!

Feb 7 2008 11:15AM

Josh Troy, the guy that runs the Stila Online business, developed an interesting VIP Card program designed primarily for customer acquisition and employee relations (friends & family).

The sturdy laminated like cards are sized just like a credit card and fit perfectly where any credit card would. The card's stylish design is clean and the VIP discount message is both clear and generous! 20% OFF and FREE Shipping. Nice!

This marketing outreach program is ideal for both new customers and employees. It's a great way to allow employees to celebrate and proudly promote the brand …and the program is a perfect way to introduce potential customers to your brand, making them feel special (like a VIP) in the process!

Here are some more 20% OFF and FREE Shipping VIP codes to be used at stilacosmetics.com.

SVPD3H2X
SVP1CDPF
SVPPVM14
SVPN7ETP
SVP4LCF4

** I'll update this list once the above have been used **

For the premium cards, you can expect to pay about .50¢ each for a run of about 1,000 cards. You can save a few bucks and use a lesser quality stock and laminate, but expect the presentation and sturdiness to be significantly less.

Where can you get these made? A lot of places, but below is the vendor Josh Troy uses.

Laura Cummings
Spectrum Digital
1707 West Magnolia Blvd.
Burbank, California 91506
818-566-4325 phone
818-566-4335 fax

Technical and Usability Tip: Be thoughtful to your shoppers and keep the code a reasonable length - less than 10 characters. Avoid characters that can be confused with like looking characters. For example, an O (oh) and 0 (zero) are often confused and could frustrate the shopper when redeeming.

A few other characters to avoid: 1, l (lowercase L), 5, S, I

Below is a SQL Server script to help you generate your own unique random codes. This code is set to create 1000 codes (change @COUNTER variable to make more or less), 8 characters in length (tweak the @code_counter ceiling to make more or less).


DECLARE @PASSCODE VARCHAR(20)
DECLARE @COUNTER INT

DECLARE @RandomID varchar(10)
DECLARE @code_counter smallint
DECLARE @RandomNumber float
DECLARE @RandomNumberInt tinyint
DECLARE @CurrentCharacter varchar(1)
DECLARE @ValidCharacters varchar(255)
DECLARE @ValidCharactersLength int

SET @ValidCharacters = 'ABCDEFGHJKLMNPQRTUVWXYZ12346789'
SET @ValidCharactersLength = len(@ValidCharacters)
SET @CurrentCharacter = ''
SET @RandomNumber = 0
SET @RandomNumberInt = 0
SET @RandomID = ''

SET @COUNTER = 1

WHILE @COUNTER < (1001)
BEGIN
	--generate code
	SET @RandomID = ''
	SET @code_counter = 1
	WHILE @code_counter < (9)
	BEGIN
		SET @RandomNumber = Rand()
	   	SET @RandomNumberInt = Convert(tinyint, 
                                  ((@ValidCharactersLength - 1) * @RandomNumber + 1))

		SELECT @CurrentCharacter = SUBSTRING(@ValidCharacters, @RandomNumberInt, 1)

		SET @code_counter = @code_counter + 1

		SET @RandomID = @RandomID + @CurrentCharacter
	END
	SELECT @PASSCODE = (@RandomID)
	
	--insert and store in your coupon table
	INSERT INTO COUPON…

	SET @COUNTER = @COUNTER + 1
	SET @PASSCODE = ''
END

Agree? Disagree? Post a comment, or email me — maybe I can add some updates to this topic if I skipped something super important.

Business rules for processing an online order

Jan 24 2008 3:44PM

I don't authorize credit cards or do address verification in the checkout process — I do that offline. The amount of effort required to contact customers due to bad expirations or a declined card is less significant than the impact you'll experience from an unreachable credit card gateway or an account issue that results in authorization interruptions. You never want to create a scenario that slows down the checkout process or possibly prevents an order from being completed.

Offline I exlucisvely use VeriSign's PayFlow Pro payment gateway — now owned by PayPal. It's easy to implement and I have found it to be an excellent, robust gateway.

Online I only do a quick Luhn check (Mod 10), which is a simple checksum formula that can be used to validate the accuracy of a credit card number.

You can't charge a customer until their order ships, except for in special cases like monogrammed or custom made products. Even if the products are drop-shipped from another vendor, you need to wait until it ships before settling. And financially you don't want to recognize that revenue until the inventory has been reduced. For online orders you need to do what's known as a '”delayed capture” transaction. You get an authorization, which is a reserve of funds for “X” days depending on the issuing bank, and then when the order ships you capture all or part of the reserved funds against that original authorization. Simple, and PayFlow Pro makes it easy.

If you're really new at this, the amount to authorize works like this….

(Product Total - Discounts) + Tax + Shipping = Order Total

Tax = Rate * (Product Total - Discounts)

(Product Total - Discounts) + (Rate * (Product Total - Discounts)) + Shipping = Order Total

Some States/Counties require you to tax shipping, so….

((Product Total + Shipping) - Discounts) + Tax = Order Total

Read more about tax in this earlier post: Tax

Address Verification (AVS) is done at the same time of authorization. The response is pretty straightforward; Y or N or X for each the billing address and zip code. X means the issuing bank doesn't support verification or unavailable at that time. The associated PNREF value to a successful authorization is key for delayed capture, just as the delayed capture PNREF value is key to issue a credit against. You can use any one of the connected to PNREF values in the PayPal manager application to see all the detail activity associated with a single order.

After the authorization and AVS process, you'll need to have some custom rules to manage orders before sending them off to your distribution center (DC). Here are few of my favorites…

1. If a customer is going to make a change to or cancel an order, its likely going to happen immediately after the order is placed, so allow orders to age at least 1 hour before their eligible to be sent to the warehouse for processing. It becomes very expensive an unproductive to ask your DC folks to track down an order once it's already in the pick-pack process!

2. Depending on the type of business and average order amount, review all new shopper high-value orders. For repeat customers that have a good track record with you, increase the threshold or bypass all together.

3. Frequency of orders should raise a flag. Review any burst of activity from a single shopper, or multiple shoppers from the same IP address.

4. Create a “never ship” list and compare your new order addresses against that list. If you get a charge back or experience fraud activity from an address, in a hash like form, store that address in a simple indexed table and compare every new address (via your hash function) against that table.

I take the first 5 ZIP and the ADDRESS, then remove all non consonants and minor address descriptors (RD, ST, AVE, STREET, CIR, etc,…)

1234 E WOODLYN AVE
PASADENA, CA 91104

Becomes…

911041234WDLYN

(you can also use this hash to find out how many duplicate shoppers you have in your database, and to manage single use promotions)

All this gets done quickly at the database level and is just part of the after order flow process. Your customer care application should have buckets where these exceptions reside.

Once an order ships you delay capture against your original authorization. If the order doesn't ship complete, you recalculate the order and capture a lesser amount against your original authorization. It gets slightly more complicated if you accept multiple payment types on a single order. In those cases you need to include a hierarchy of payment distribution.

Agree? Disagree? Post a comment, or email me — maybe I can add some updates to this topic if I skipped something super important.

Site Conversion
(even if you don't read this, I'm counting you)

Jan 19 2008 7:10PM

Conversion is generally measured and reported as “Visitors to Buyers." However, most managers that have the responsibility of running and building an e-commerce business rely on some other site conversion metrics as well.

“Items in Baskets” versus “Items in Orders” is a good metric to know - not to be confused with “Shoppers to Orders,” which although important isn't the same thing. For example, I can add 5 items to my shopping basket, but end up only buying 2. A shopper to order metric would show this as 1:1, when really I bought less than my original basket value (5:3). Site managers and merchants should be interested in what 3 items I abandoned. Maybe there was an unanticipated surcharge in checkout that changed my mind?

I have seen a lot of sites that create baskets for the visitor on the first page load. The retailer sometimes will count them as a shopper, but they really aren't at that point. Online developers do this because there are often attributes to that initial landing that they want to associate with a potential resulting order (site referral links, click through discount offers, etc,…). But the visitor isn't really a shopper until they show some sort of shopping behavior, like adding an item to their basket.

Actual "Shoppers to Buyers" conversion will always be higher than “Visitors to Buyers.” Sites that measure conversion based solely on the first usually have over-inflated conversion metrics.

Individual item (SKU) conversion is an important metric, too. To properly track this you need to correlate detailed inventory data to those items for the same time period. Here is how it works….

A parent product or style can be fully in stock, out of stock, or partially in stock. The first two need no explanation. Partially in stock simply means that at least one of the items under the parent product/style is out of stock. For example, this red shirt is only available in small and large, medium is out of stock. Maybe the most popular size is out of stock, which of course will impact that item's conversion rate.

So this means that every time a product detail page is viewed, you log the inventory position of that parent item, by unique visitor. Once you have some good history, you can easily measure the impact on sales/conversion for items that aren't fully in stock.

Related, there is usually a silent struggle between merchants and site managers when it comes to product statuses. In the example above, a merchant/planner/buyer might want to inactivate the broken SKU or parent product — meaning remove it from the site completely. If the SKU is coming back in stock, a good site manager will want to keep it active and measure some of the missed demand. They may even try to recoup some of those misses by having a “notify when back in stock” option.

It's nearly impossible to find reliable conversion benchmarks — it's always apples to oranges since everybody slants their numbers in a variety of ways. There really isn't a standard you can look to because aside from each vertical being slightly different, traffic patterns, merchandising and promotion strategies vary wildly. A company that gets little traffic, but is overly promotional will always see higher conversion rates. And new traffic usual converts much lower rate than your repeat purchasers. Yada, yada, yada….

Your best shot to understanding your site traffic/conversion and customer shopping behavior is to have mix if in-house data collection and a 3rd party analytics provider.

The best bargain out there is Google analytics, which is free — but you do get what you pay for. It's a nice superficial overview of your online activity, but it lacks the robust deep-dive reporting and analytics you can get out of an Omniture, Coremetrics or Visual Sciences (formerly known as Websidestory and currently being acquired by Omniture). For a company with limited resources — Google would make a lot of sense.

Omniture is the best. amazon's ECS and GSI's and MarketLive's platform all have a default analytics offering of Omniture's services, so you know it measures up to their high standards. I've used Coremetrics but never could get their collection data to match up perfectly with my own. Omniture is pretty much spot on, and my experience with them has been great. However, the reporting suite/portal can be a little confusing - imagine a system that was created by really smart people, and they assume that everybody using it is super smart too. But now that they'll own Websidestory (my second favorite), maybe things will get a little moreuser-friendly.

Agree? Disagree? Post a comment, or email me — maybe I can add some updates to this topic if I skipped something super important.

I'm waiting for Bill Me Never

Jan 14 2008 7:41AM

As you may already know, amazon recently made an investment in Bill Me Later, as well as committed to offering them as a payment option in checkout. I sort of remember amazon having Bill Me Later on their site many years ago… does anyone else remember this? Why did they take it off their site for so many years?

Bill Me Later is a secure deferred payment service that allows consumers to pay for purchases up to several months later. It's basically another financing option for consumers (like a credit cards, or PayPal's Pay Later service).

Here are a few reasons I like Bill Me Later...

1. Why not give your customers as many payment options as possible? As long as those payment options aren't of the likes of FLOOZ or BEENZ dollars. Whoopi?

Seriously, offering reputable alternative payment options like Bill Me Later and PayPal, along with your traditional credit cards, is just thoughtful for your customer. You're leaving money on the table if you don't... and if you think it might hurt your brand, you're probably wrong.

2. Bill Me Later allows the retailer to do some interesting (creative) promotions. For example, no payments for 90 days, or 90 days same as cash, etc,.... I make good $ and always use the deferred financing options when buying appliances or furniture.

3. You'll see higher average order sizes using that payment type, but I doubt offering Bill Me Later is going to result in a significant amount of new traffic. At the Bill Me Later site, they do have a large merchant directory (and growing), and I suspect a few people that are stretched financially thin might look to the “Bill me Later” merchant network for shopping options.

However, there is some real downside to offering Bill Me Later as a payment option. Bill Me Later requires a better than average credit response for approval. In fact, on average I think there is a 50% rejection rate (ouch!). So if I load up my cart - thinking I can pay for this stuff later - then get denied in checkout, well, I likely won't buy anything at that time. I'll take my poor credit score (and what's left of my dignity) and abandon my cart.

Also, when I met with Bill Me Later years ago they really wanted me to over brand my site with the Bill Me Later payment logo (home page, product detail page, checkout, etc,...). They told me if I didn't then I wouldn't see the strong results many of their other customers were seeing. It seems odd to me to promote their payment option on products other than the high-ticket items or larger than average carts.

Check them out: Bill Me later

Agree? Disagree? Post a comment, or email me — maybe I can add some updates to this topic if I skipped something super important.

Victoria's Big Secret? No gift giving!

Jan 7 2008 7:16PM

I can understand a retail site not offering gift wrap on orders, but they almost all give you the option to suppress pricing on the packing slip so the recipient doesn't see what you paid and all the deep coupon discounts and credits you used on the order. How does a site like Victoria's Secret — one with a ton of gifting potential — not offer this option in checkout?

And even if NO pricing is printed on packing slips (gift or otherwise), which I suspect is the case, then why do they not mention that in checkout? They're definitely leaving money on the table. And wouldn't they want to know what segment of their shoppers are gift givers, which you can't always get from just differing billing and shipping addresses?

One more thing, why do they make you enter zip code when logging in? Whatever you think the reason is, it's not a good one.

Agree? Disagree? Post a comment, or email me — maybe I can add some updates to this topic if I skipped something super important.

Tax

Jan 4 2008 8:37PM

If a tax Nexus applies, meaning you (the tax payer) meet the minimum requirement where the State can impose a tax for your commerce activity, you'll need to figure out how to handle your sales tax. If you're a resident company, have stores, employees, a warehouse or some other physical facility in the State, chances are a tax nexus applies to your business.

Here are a few points about sales tax….

Sales tax applies to the State where the product is being shipped, not billed.
Tax rate varies by county/zip code. For example, in California, the sales tax rate in Pasadena is 8.25%, but in San Diego it's 7.75%
Some States require that you collect tax on shipping and/or special handling (by county).
Some States has some complicated product tax hierarchy rules, meaning certain categories of products have a different tax rule than other categories.

For businesses that have products that span multiple categories, you definitely need to consult with a tax professional. However, if your product offering is pretty straightforward, and you have a financial resource that's comfortable with the tax filing process, then Tax Data Systems might be something to seriously consider. For under $1K, annually, they give you a monthly updated dataset (flat file) of all the tax rates, by CITY, STATE, ZIP, COUNTY for the US and CANADA. The simple file even indicates if the shipping and/or handling should be taxed.

Check them out at http://taxdatasystems.com/

From a developer's point of view, nothing could be easier! One simple table.

Your finance resource will still need to manage the tax forms and comply with the filing guidelines, but the Tax Data Systems is a good setup and it's going to make you a lot more compliant than most folks out there.

Keep in mind, there are several solutions and services out there if you have too much money and scare easily... and have a lazy finance department.

You may have heard of the voluntary Streamlined Sales Tax Project (SSTP), which among other things is setup so that non resident companies can voluntarily pay taxes to the states where product is shipped from their e-commerce activity. Why would anyone voluntarily collect tax from their customers when they may not have to? Not sure, but talk to a tax professional if this topic is keeping you up at night.

Agree? Disagree? Post a comment, or email me — maybe I can add some updates to this topic if I skipped something super important.

Inventory - Oh, the Humanity!

Dec 28 2007 8:36PM

Yes, a mind-numbing topic! However, good inventory management is a foundational component to successfully managing any retail business. "Master of the obvious" stuff, but I'm going to say it anyways....

Your wholesale/retail customers will hold you to high fill rates. On the direct to consumer side (web store), short ships create more work throughout departments, and the unexpected online experience frustrates customers and greatly reduces repeat purchases. And since a retailer’s borrowing base is usually linked to the inventory they own, your finance folks probably require accurate inventory visibility. And Planners? Good inventory management is the horse that needs to come before that cart.

The best way to manage inventory is to tightly map your Warehouse Management System (WMS) transactions to your Financial/Inventory system. All shipments, receipts (purchase orders and returns), miscellaneous inventory adjustments and "cared about" movements need to be interfaced so you know what inventory you own is available to sell and what exists in non-sellable locations, by non-sellable location.

Mapping those transactions is an important exercise that requires dedicated time and input from the Opertaions/Logisitcs, IT and Finance folks. For the most part, shipments and receipts are fairly straightforward, but the movements and adjustments can be pretty confusing with different reason codes and accounts that those transactions need to hit. You may not need to map sellable to sellable movements, but you'll want to map non-sellable to non-sellable movements. Why? For example, you need to be able to differentiate (and interface) what's moving from damaged to return to vendor, versus damaged to rework.

You should get an end of day location based inventory snapshot from your WMS for a few reasons.

1. Your financial/inventory/ERP system will handle inventory like a stock ledger;

beginning inventory + receipts – shipments + or - adjustments = ending inventory.

But your DTC web store might be better off having inventory replaced nightly based on what your WMS snapshot reports as available to promise/sale (ATP). ATP inventory is simply your sellable (good) physical inventory minus all allocations. An allocation is demand against the ATP inventory that hasn’t yet shipped.

Using the nightly WMS inventory snapshot, an inventory replacement formula for your web site would look something like this;

ATP - DTC sold but not dropped = site inventory.

If you're sophisticated enough to know what other non DTC demand has been promised but not dropped and allocated, then you can expand that formula to include that inventory;

ATP - (DTC sold but not dropped + non DTC demand) = site inventory.

2. A good inventory manager addresses and works inventory discrepancies daily. With an end of day location based inventory snapshot from your WMS, you’ll be able to create delta reports against your financial/inventory/ERP system that highlight those discrepancies. Once you're interfacing miscellaneous transactions, movements, receipts and shipments, your inventory manager will be able to quickly identify transactions not mapped or missing. Oh, and a good inventory manager needs to maintain a good relationship with the warehouse crew!

Many big retailers have multiple warehouse location, often splitting the DTC fulfillment from the store and/or wholesale distribution. That may sound more complicated, and it is for the planning and allocation teams, but from an IT and Operational point of view this can be a nice approach. I have found that most warehouses that are setup to fulfill retail and wholesale channels often fall short on the small pick and pack needs for a large DTC business. And vice versa!

I could talk about inventory all day, and the above really just touches on a few points. Post a comment, or email me — maybe I can add some updates to this topic if I skipped something super important.

Serious Retail Technology and More

A picture is worth a thousand words.... and a few dollars!

How can you not take credit cards at a trade show?

Pedestrian Counters

URL Rewriting

How to minimize online fraud and abuse of marketing promotions

Google is one very costly spell checker

Secrets of Online Conversion

VIP Cards - very smart!

Business rules for processing an online order

Site Conversion (even if you don't read this, I'm counting you)

I'm waiting for Bill Me Never

Victoria's Big Secret? No gift giving!

Tax

Inventory - Oh, the Humanity!

Site Conversion
(even if you don't read this, I'm counting you)