Could Your Business Processes Cope With a 500% Increase in Volume?

Today we’re wrapping up The Tapas Lunch Company for another year. The final orders went out to our customers last night, and although we’re open today, hopefully we’ll be sat with our feet up eating mince pies (or the Spanish equivalent) most of the day. Personally, I’m excited because this evening we’re going to drive down to Madrid and on Saturday we’re jetting off to (marginally) warmer climes to spend a couple of weeks with family – a perfect, relaxing end to a hectic couple of months.

And a hectic couple of months it really has been. Tonight we’ll know exactly by how much we’ve beaten our records, but we already know we’ve beaten them for a second month in a row. Christmas, naturally, is always busy for us as an online retailer. The rush starts around mid-November and builds to a fever pitch around the second weekend of December. Trade customers are very active in November, preparing for the festive season, so the increase is more in turnover rather than number of orders. In December, as private customers shopping for gifts and treats for their Christmas table visit us, the number of small orders shoots up.

We’re familiar with this pattern, but this year the volume caught us by surprise. We were already on a decent upwards growth path this year, so the boost in sales this December has been far more than we expected – roughly double last December. And December is already roughly doubly as busy as ‘normal’ months. That’s a four-times increase in monthly trade. And seeing as last orders were yesterday, the 21st, that makes December roughly only two thirds the length of other trading months, so the effective multiplier for the volume of work is more like 6 than 4.

What did we learn from a 500% increase in volume?

Well, basically, that our systems are crap.

Well, perhaps not crap, but they’re certainly not as slick as they should be. They work most of the time, but when exceptions occur and humans (me) have to get involved, errors happen. And more than I thought.

An ‘exception’ is some event in the flow of a transaction that is outside the norm. Depending on how a business runs, this could be pretty much anything that separates a standard order from a non-standard order.

I actually have a script that runs on all incoming orders and flags any that have particular characteristics. That way, 80% of all orders just flow naturally from the customer’s basket, through the ERP system, through the warehouse and out to be despatched. The 20% minority that need some form of human inspection are flagged as such and I can quickly run down the list and see if anything is amiss. There aren’t that many reasons (thankfully) why an order would need my attention; here are some examples:

• Customers who live in a remote area and haven’t paid extra for delivery.
• Wholesale customers who haven’t conformed to the minimum order requirement.
• Customers who have made some special request regarding the delivery.
• Customers who have selected PayPal and not paid.

It’s a pretty efficient system for dealing with non-standard orders and I’ve always considered that this script, combined with our other business processes, produces only a negligible number of what you would call errors, even confronted with orders that are somehow out of the ordinary.

A good definition of negligible would be that even when amplified by 500%, the number of errors would still be, well, negligible. And that isn’t what happened. Instead, I was confronted with the sobering reality that our business could not scale by 500% without some serious work to the fundamental mechanics.

And that’s not acceptable to me.

A business should operate as a perfectly oiled machine, happy at 30mph, but just as comfortable at 120mph.

The ins and outs are probably not that interesting, but just to give you a flavour of the mild chaos that we witnessed, I’ll give you a few examples.

Please Despatch on the 21st

Normally orders are despatched the day they come in. From time to time, customers write a little note in the ‘Delivery Instructions’ field that says something like ‘Despatch on the 21st’. This is rare, so we don’t have a system for it. Instead, that order just sits at the top of the list, and it’s up to me to notice and release it on the day it’s due to be despatched. That’s fine when there are one or two on the list, but when customers start placing orders in November that are due to be despatched just before Christmas, it causes mayhem. It got to the stage, in December, where the first 60 orders on my list were for future despatch dates. It was too much to manage and led to many errors.

Please Put One Aside for Me

We don’t normally allow customers to back order. If we’ve got it, we’ve got it, and if we don’t, we don’t. But we had a couple of really popular products this year that we simply couldn’t get enough of. Customers were calling asking if they could reserve one on the next incoming batch. At first I told them that they couldn’t, but it felt like I was just going to provoke a free-for-all the day they came in. So we turned on backordering (just for those products) on the website and allowed people to reserve them. Again, we hadn’t done this before, so we didn’t have a system. I had to try to keep pending orders separate from current orders using a manual system like the one above and it was disastrous. To make things worse, customers were combining the two issues, reserving stock and specifying despatch dates later than a month away. Before we knew it we were selling stock that we didn’t have to people who would be expecting it two days before Christmas. It was horrible to see what I have always considered to be such a smoothly run process descend into a farce. We pissed a lot of people off.

Couriers.  The Bane of My Life

It’s no secret that all e-commerce outfits have issues with deliveries – it goes with the territory. But as the volume went up, I was struck not just by the quantity, but also the staggering variety of issues that came up: delays, failed deliveries, cards that were never left, recorded deliveries that never actually happened, damages – whatever you can imagine, our courier managed to do. I’ve been doing this for years and thought I’d seen every possible courier mess-up. I hadn’t.  I think, over the years, that we’ve tried all the couriers out there, but I’m willing to keep trying.  I’m convinced there must be better out there.

The Dreaded Blue Screen

Like every business, we have the odd technical failure. When a script fails and you have to manually enter 20 orders, it’s frustrating. When there are 100 orders to be done, it’s a virtual impossibility.  I never thought twice about little computing glitches like Javascript errors or bugs on the website.  They were so few and far between that I could jump in and patch them whenever necessary.

Tap, Tap, Tap

Similarly, our Amazon and Ebay sales really took off this year, which meant having to manually enter order data for a couple of hours a day. We’ve been selling on Amazon and Ebay for a while, so we’re used to entering the odd order. But before, it somehow never felt like a drain on resources. You took five minutes to do it between other tasks. But this Christmas it became a serious chunk of time that had to blocked out. To me, manual data reentry is a total waste of time.

The truth is, you never know how your business systems will survive until you test them at their very limits. What seemed sporadic, negligible failures and errors before turn into a systematic drain on quality and efficiency when multiplied by such a large increase in volumes.

I’m a little obsessive over systems and automation of tasks where possible, despite the small size of our business. I want everything to run exactly as it should and my aim is to have the number of ‘exceptions’ reduced to such a level that the business can scale almost infinitely without the need to either compromise quality or patch systems over.

A lot of traditionalists would probably suggest that we should have employed someone to deal with this increase in volume – and that’s the exact opposite to my way of thinking. Paying people to do jobs that automated systems should be doing is not only a waste of money, it’s also a surefire way to create the most mind-numbing, spirit-sapping sort of automaton job that exists. I wouldn’t be doing the business a favour and I certainly wouldn’t be doing the potential employee a favour. People are supposed to run efficient systems, not act as crutches for rubbish systems.

So next year, rather than putting a suit on and going out schmoozing billion pound multinationals, or visiting trade shows tasting £30/litre olive oils, I’ll be doing the dirty work of redesigning some of our fundamental processes, reprogramming our IT system, changing our courier and laying the groundwork for a better business. I believe, in business, that you have to be the absolute best you can be. I also think this is what customers want. I’m just glad that it was a fantastic boost in sales rather than bankruptcy that gave me a kick up the entrepreneurial arse.