Question Week: Question #2

Here we go:  "Kevin, how should I evaluate the productivity of an item we sell? All of our reporting is channel-centric. How do I obtain a complete view of the performance of an item? Thanks, Sherri."

Good question, Sherri!

I'll share an answer with you. Not one soul is going to agree with me. Not one. Everybody will have their own ideas. Few are implementing their ideas. So let's at least get some thoughts going.

My recommendation is to evaluate item-level performance on a monthly basis. The image below is for twelve months. One can compare the current twelve-month period to a prior twelve-month period ... I'm not going to go into that level of detail here, as the concepts matter more than the comparisons.

Here's a potential template - click on it as the text is tiny, peeps.

Ok, let's explore sections of the analysis.

The top set of rows illustrate where sales come from, be it a physical channel like a call center, catalog orders matched back to the online channel, online sales by online marketing tactic, all of it rolls up to what I call Gross Demand. This is what the customer "wanted".

Then our leaky bucket subtracts some of what the customer wanted. Items that were not fulfilled and items that were returned are subtracted from Gross Demand, leaving us with Net Sales. We have units sold, and a simple division yields price per item purchased, an important metric.

The next section deviates from sales for a moment. Here we analyze how many visits the item attracted online, how many orders were generated for the item based on who viewed the item, and the conversion rate for customers viewing the item.

Here's the part you won't like. It's hard to generate this information at an item level, but it is absolutely necessary. On a monthly basis, you sum ad costs by marketing channel ... at an item level. Yup, not easy, and you'll come up with a thousand reasons why you shouldn't do this (but the real reason is because it is hard to do it). If you have a catalog, allocate square inches against your life-of-catalog curves (i.e. 40% of sales happen in week one, 30% in week two, 20% in week three, 10% in week four), yielding an ad cost for that specific item.

Digitally, this becomes interesting as well. If a customer visited from search and looked at three items, the cost of that click is allocated to the three items the customer looked at. If a customer visited your home page and bounced, you allocate that visit across all items equally (I know, that can be punitive, but you'll view your assortment differently after doing this). Sum up all ad costs, and you've got the amount of marketing dollars spent against that item for that month.

That's a big deal, folks!

Sum up your pick/pack/ship expenses for that item next. This gets you to what I call "Variable Profit" ... it's the amount of profit the item generates prior to fixed costs. Fixed costs don't change as sales change, so you don't evaluate fixed costs in this framework.

Again, this style of analysis requires hard work and discipline, so most readers aren't going to pursue this path.

If you want to know how profitable each item is, pursue this path, ok?

The next three rows evaluate gross margin dollars and rates. Obviously this is a big deal.