Sep 82009

Keep it Complicated

How many orders did we take last week?

How could such a simple question start so many board room arguments? Yet it does, and not because companies don't know their numbers, and not because companies lack a unified dashboard with definitive answers. The reason is: the answer isn't simple.

Considerations when counting orders:

  • Do you count orders that are placed but not yet shipped?
  • What about completed orders that are subsequently canceled?
  • Do you count back orders?
  • If someone calls back and adds on to an order is that one order or two?
  • Do you count orders made on credit or installments before they're paid in full or after?
  • How do you count returns?

Your answers to these questions will depend on your role in the company. The operations folks are interested in call volume, and boxes shipped; the accounting folks will focus on money in the bank; the merchandising folks and marketing folks have different facets of the business to tune and must look at these numbers differently.

I'm not saying that every discussion needs to be down in the weeds of the details. Arguing over whose numbers are "right" is a waste of time. There are plenty of instances when agreeing on the definitions in advance makes sense to simplify higher-level discussions. I am saying that this shouldn't be the only way of looking at the data. Each department needs to focus on its own customized view of the business, not to the exclusion of other perspectives but to enhance their understanding.

Simplifying and unifying the company's view of numbers is dangerous. There are good reasons to look at the numbers from different angles and with different agendas.

I use order counts as one example. The same notion applies to each marketing program and the inter-relationship between them. The unified view is important, but it shouldn't be the only view.

Business is complex. Reducing the complexity means losing valuable information. Sweeping the details under the rug of a dashboard serves no one's interests.


6 Responses to "Keep it Complicated"
Jim Novo says:
Likewise, "How many customers do we have?" Include Inquiries? Ordered but not paid for? (check payment) Cancelled orders? All orders returned? 1x buyers? And of course, my favorite: Last interaction was more than 2 years ago? More than 5 years ago? More than 10 years ago?
Marc Adelman says:
George, Your post got me thinking - or rather resurrected an old thought - or line of questioning if you will. When it comes to the value of a converting KW - should we discount complete value for an order that was canceled, returned, not yet processed or shipped, or any of the other examples that you or Jim mentioned? Do these post order activities actually add or detract from the value of the KW and its tendency to convert? I often think that these other post order activities have little connection to the kw that lead to the conversion. For instance, can I really say that one kw leads to more canceled orders or returns, or does that have more to do with everything else in the converting process from landing page to cart experience and even beyond to fulfillment and actual product related issues. If one were to run bid rules on KWs attributed with orders excluding returns and cancels, would it be possible that positive performing KWs could receive a bid change that would decrease impressions - clicks - and conversions? On the flip side - would it be ludicrous to credit all orders regardless of current status "the orders was placed - end of story", to allow KWs that lead to conversion (regardless of post order activity) to have a chance in leading to more conversions? * All of that being said - If there WAS a pattern of cancellations or returns for any given KW - there could be a flag for that - as it would still point more to a on site issue, but until that was resolved - it might be a good idea to change the bid or pause.
David says:
Hi George, long time reader, first time writer. I couldn't agree more, it's so important to have a consistent definition of what "orders" are to ensure everybody's on the same page. Consider this: I recently talked to the CEO of an SEM agency, who claimed that his company had been able to increase sales by 140% while reducing costs by 60% since taking over the client's account. Knowing which agency this partcicular client was with before, I was very sceptical. Solution: What he was talking about was "# of items purchased" vs. "# of orders". So basically, when a customer purchases two pairs of socjs, that's 2 Orders! It hadn't been like that before.. Also, it didn't help his credibility that they now use a 365-day cookie compared to a 30-day cookie before. Defining what is what in search is so critical!
Thanks for your comments, gentlemen. Jim, the notion that some "customers" might actually have negative value is a subtle but important one. Particularly when marketing to customers is direct and expensive (direct mail), measuring the real responsiveness of customers whose first purchase was returned entirely is a valuable exercise. Marc, your point is well-taken. The flip-side of Jim's. Does the fact that a customer returns an order make the KW that generated the initial order less valuable? So much of the most granular data is one-off that the most atomic view is essentially impossible. What if that one order attributed to a KW happened to be very small, or very low margin, or (as is sometimes the case) negative margin? Figuring out what's the signal and what's the noise is both stats and art -- and guesswork. David, we've seen this happen as well. There almost needs to be a check-list associated with any such claims to get to the real truth. George


Check out what others are saying...
[...] Keep it Complicated – view page – cached Simple isn't always better. — From the page [...]
[...] Keep it Complicated, Rimm Kaufman [...]

Leave A Comment