What’s your Penny Rounding Story?


If you've been around the Order Management System (OMS) space long enough, you likely have a good penny rounding story to share with your peers. It's a battle scar we proudly bear as OMS veterans.  

At this point you are either smiling because you know exactly what we’re talking about or you are wondering why in the world Nextuple is asking you about your penny rounding story.

It goes like this...

“I remember back in 2014 we integrated our OMS with our new (insert your favorite tax engine here) system. We all went home (because we worked in real offices back then!) because everything was working as planned. Heck, even the tax guy was happy! But it was Sunday afternoon when the help desk calls started. Customers were calling asking why they were being charged for one penny! The call center was being overwhelmed. It was all hands on deck and we had to figure out how to refund 5000 customers for one lousy cent!”  

Perhaps it was a new ecommerce engine being hooked up that rounded promotions to line differently than the OMS did? Somewhere, somehow you were burned by this. And if you haven’t been, it’s just a matter of time.  

As the orchestration engine across multiple systems responsible for authorizing and settling payments, registering taxes, etc - the OMS is typically in the middle of a failure to understand the intricate rules of a system’s rounding behaviors when it comes to financial transactions. And it usually takes the edge cases of production level data to produce these rounding errors.  

In the world of order management, things will go wrong no matter what your QA process looks like. If not penny rounding, it might be special characters between two integrations causing them to fail. The list goes on and in today’s environment of fast shipping and BOPIS order ready times measured in minutes, a retailer can’t afford these mistakes in production.  

The solution?

We recommend a “belt and suspender” approach to order monitoring to quickly find these kinds of errors in real time. Point to point integration monitoring could be the belt.   But monitoring like this needs business context. If the Order Management System is successfully authorizing but collecting one cent, everything looks fine. The “suspenders” is the monitoring of your entire business process against expected business service levels and historic baselines. For example, if the number of authorization attempts is spiking in the OMS against baselines, this demands immediate attention.

Typically we see retailers react to these issues with point solutions that solve the specific issue at hand. They may deploy an order reconciliation handshake between two integrations looking for a specific functional issue and correcting it.  

A better solution is to think about this holistically and to realize that no matter what you do in QA, something is going to go bump in the night that you didn’t plan for.

A capability that has the ability to monitor your OMS business processes flows by looking at event data measured against historical trends and service level agreements is designed to catch anything that isn’t moving through that process as planned. Order management vendors solve for this with their own alerting tools.

These solutions put unnecessary load on your transactional system and we typically see the alert fire down at order line level. This can create lots of noise in your monitoring solution. A monitoring solution needs to group like issues into alert events which can then integrate with your incident management system.

There are solutions in the marketplace that attempt to solve this. We’ve worked with these systems but what we find is that they lack the domain context and require extensive setup against the OMS data model to begin adding value. There are large enterprise control tower applications that are better suited for inbound supply chain event management vs. consumer facing fulfillment. 

At Nextuple we support multiple OMS customers through managed services with our own purpose-built OMS monitoring solution, OMNI Control Center. It comes with an adapter you can quickly install that reads events out of your OMS and turns them into a real time dashboard of your OMS health. And because it can read any OMS event, you could easily deploy it to solve business monitoring use cases like order ready times at stores or late shipments from drop ship vendors.

Wherever you are in your journey to scale order management, a robust monitoring system is a must to mitigate ballooning manual support costs and more importantly, to catch that little penny rounding error in its tracks before it costs you a whole lot more.  

New call-to-action