### Saks and Robotics: A 21% Sales Increase?

You probably read the article in the NRF SmartBrief about how robotics helped Saks achieve a huge sales increase via robotics.

There is one sentence in the article where a link between sales and robotics is made:
• It seemed to make a difference: Saks Direct reported a 21 percent increase in fourth quarter sales from the same period in 2010.
What fraction of the 21% increase in e-commerce sales would you attribute to a warehouse robotics system?

Well, you might look back to Q4 of the prior year to see what happened.  In this article, we learn that Saks e-commerce division posted a 36% increase in Q4 of 2010.

So this tells us that e-commerce was on a major updraft in 2010, and the rate of improvement slowed in 2011.  Regardless, there is momentum, momentum we cannot attribute to warehouse improvements.

On Twitter, @richardfergie suggests that customers who know they can order up to December 23 will spend more than customers who lack confidence that their order may not arrive in time for Christmas.  Ok, this is a good hypothesis!  How would we test it?
• We could have offered up a messaging test ... 80% of the audience is told that they can order up to 12/23, 20% of the audience is told that they can order up to 12/18 or whatever the old date was ... then measure the incremental lift between the two groups.  This would tell us what impact the robotics system was likely to have.
• Analyze the distribution of orders in 2010 and 2011, to see how many orders were pushed late (i.e. after 12/18), then give credit for those orders to the robotics system, assuming that these orders would not have happened otherwise.  In other words, if a customer was going to order on 12/12, then orders on 12/19 because of a promise of delivery by 12/23, the net financial impact is \$0.  If a customer was not going to order, then orders on 12/19 because of a promise of delivery by 12/23, then the robotics system (and marketing messaging) get credit.  If the customer was going to spend \$100, then elects to spend \$200 because of a promise of delivery by 12/23, then the robotics system (and marketing messaging) get credit.
Now, via analytics and testing, we might prove that a robotics system more than pays for itself. I'll bet somebody in Finance / Operations would love to have an estimate (and it would be an estimate).  That would be a good thing!

We can also measure how much more accurate pick/pack/ship activities are when using a robotics system.  Knowing the improvement in accuracy, we can calculate a cost savings per order, multiplied by annual orders, yielding annual cost improvements.

We have the tools and techniques (you probably have more ideas for how to analyze this than I have, offer them up in the comments section) to measure the impact of a robotics system on the profit and loss statement.  Let's go measure it!