Attempting to Quantify How Busy We Are

Detailed Image has had a really busy Spring season. Everyone has had to chip in a bit extra at the warehouse to ensure that we get all of the orders out the door in a timely manner. This past week was so busy that Mike, Greg, and I went in Sunday to pack boxes for about five hours, something we’ve never done before outside of the Black Friday to Cyber Monday rush. We’ve also each had many days where we’ve had to go into the warehouse unplanned to help out.

One of the more difficult things has always been figuring out at what threshold we need extra people in the warehouse beyond the scheduled team. What we’re really trying to quantify is: how much physical work needs to be done to get all of the orders out? For years we always relied on the total dollar amount of orders in our shipping queue. This is a good metric, but it has some noticeable flaws, especially on busy days. For instance, if we’re running a 20% off sale, that number will appear to be lower than the actual amount of work that needs to be done (an order that usually costs $100 will only be factored in at $80). Or, if we have a buffer on sale (as we do this month) and we’re selling a lot of them, it could make it appear as if there’s more work than there actually is because there’s a big difference in how long it takes to package a single $130 item than $130 in $10 – $20 products.

Given this, we asked ourselves, what numbers do we have that will help normalize this data a bit and give us a simple number that we can work off of on a daily basis? We decided that we could get a much better idea of how busy we were if we factored in total dollar amount being shipped, total number of packages being shipped, total number of items being shipped, and total weight being shipped. We then came up with a maximum threshold for each, the absolute most we think we can do in one business day given our current team, which we guessed to be roughly 20% higher than our previous maximums. These maximums will likely change any time we add staff or introduce a new efficiency.

We weighted each of these factors equally to come up with a simple single percentage that gives us a better approximation of how busy we are. We can view it in “real-time” for what is currently in our shipping queue, and also have a report so that we can compare it to the past. Our maximum is roughly five times larger than our average day, meaning that on average we’re around 20% capacity but that we’ve fluctuated up to about 80% on our busiest day. I’m not sure if this is normal for e-commerce companies or not. I certainly think the seasonal nature of our business, combined with how often we run sales, factors in to the variation quite a bit.

Here’s a shot of the variation from before, during, and after a recent sale (read from bottom up, so the 15% day is the first day in the example):

Busy metric

I’m really excited about this. It’s a very simple metric, which is why I think it works. It’s hard to try to think in terms of packages and revenue relative to our staffing capabilities, but it’s easy to think in terms of percentages. In both the short term and long term this should help us out quite a bit with our warehouse staffing decisions.

6 comments on Attempting to Quantify How Busy We Are

  1. Tim says:

    Very cool, I like simple solutions like this! It’ll be cool to see how accurate this is for you as it is tested over time.

    • Adam McFarland says:

      Thanks Tim. The hope is that we can get it to a point where we can have set percentages where certain things happen, like someone coming in early or an extra person coming in (possibly even alert those people with a text or email). I’m sure there will be some tweaking involved, and there is some ambiguity in the maximum values for sure, but all in all I think it will be very useful for us.

  2. Rob says:

    Ooh, interesting idea. Do you know things like your average order pack time for all products/specific products and mean, standard deviation and so on? I’d think with that information you could also build quite a good picture of how much work there is to do in terms of time at least. I can’t imagine that dollar value is particularly accurate considering it could be many small items or one large order etc.

    • Adam McFarland says:

      Do you know things like your average order pack time for all products/specific products and mean, standard deviation and so on?

      Unfortunately no. That’s the data that could really tie all of this together, because then you’re able to calculate a true maximum, and you’re able to say how long it will take you to pack on any given day. Right now we’re just taking an educated guess at the maximum amount that we can pack in a day. We’ve discussed the possibility of collecting this data, but it would be a lot of work and there’s a ton of noise. For instance if a delivery comes the guys will stop and receive it, unload the dock, and then get back to packing. That needs to be factored out. In the future if we have security cameras it could be much easier to track. I could definitely see myself spending a day collecting and analyzing times if that ever happens.

      I can’t imagine that dollar value is particularly accurate considering it could be many small items or one large order etc

      It’s an OK measure. To some extent those things average out. It was good enough to get us through the past few years, but it certainly has it’s problems. It seems a little more appropriate to have it factor in 25% like we do now.

      Ironically, a week in to this, I think the harder thing will be changing our habit of referring to everything in terms of dollars. Every time it’s come up we’re still saying “there are $x in queue, we better get out there” and not “27% in queue” because we all think in dollars and not this new percent. It will take a little time…and I’ll have to make sure I don’t revert back to old habits.

      • Rob says:

        Why do you need to factor out receiving a delivery? If that’s a real situation and happens often, shouldn’t it be left in? After all, that would give a more true to life picture of what’s going on.

        I think you’d need to collect a lot of data over a lot of time to get a clear picture – if people knew it was just happening on a specific day they might work to a different pace. If you had some way of logging when an order started to be picked and when it completed packing, combined with the order details, you could get an awesome data set. However, to get a system that can log all of that might be more effort than it’s worth – you’d need to add things like a recording method (barcode, RFID or just a computer at each end of the chain where people could check in) and of course this would mean you’d have to spend time doing this… recording activities on camera is undoubtedly less invasive to your current processes and simpler to implement but would require more time spent analysing data and might not be able to show what items are in the order. A very blunt method to get an overall average would just be to “clock in” when you start doing orders, “clock out” when you’re finished and work out the mean time per order that way. Very very blunt though!

        I guess if you’d got used to doing a dollar value then it’ll take a while to change your way of thinking of it. Did just use the number of orders/items in the queue? eg. “we’ve got 40 orders to go out this morning”?

        • Adam McFarland says:

          That’s literally the exact thought process we’ve gone through Rob. A barcode system would be huge for data collection and the elimination of errors, and we’ve discussed it a few times, but the conclusion we’ve always come up with is we’d probably need to be 5-10 times larger to make it worth our while to implement and maintain.

          The problem with deliveries is that they can take 30 minutes to unload (not even to check and unpack) because of our small dock and lack of a forklift. We sometimes need to literally take a 1,500 lb order off of a truck box by box, and then move it box by box to a cart to get it out of the way for the next delivery.

          Many days this doesn’t happen at all, other days it can happen five times, which would really throw the data off. Now, if we were collecting a ton of data like you suggested, it would all even out over time. But the more likely scenario is that we’ll have cameras within a year or two and we’ll be able to take a week worth of data to get an estimate, in which case we’ll have to factor out things like deliveries and breaks.

          We did also used to reference the number of orders going out. For one reason or another, we slowly migrated to dollar values and that’s how we’ve been referring to everything the past year or two. Overall though, I think number of orders and number of items are both better indicators.

Comments are closed for this post.