A Business Case for Price Optimization

My first price optimization implementation yielded $45 million in profit in three months and forever changed my views on pricing software and benefits to this day.

Several years ago, I supported an international US-based retailer. During this period in my career, I implemented large retail ERP systems consisting of multi-year, multi-phase projects, detailed hardware sizing, hundreds of integration points, and countless hours of process reconfiguration, data transfers, training, and project management, all without disrupting the retailer’s day-to-day operations. When implementing an ERP at a large retailer, your job is picking up the Empire State Building, changing the foundation underneath, and then gently setting it back down.

What made price optimization different? Why did it work? And what could have been better?

 

It allowed us to find opportunities.

A small price change, either upward or downward, can significantly impact the overall sales of that product and complementary products. That impact varies by item, as some items are more price-sensitive than others. Finding where to change the price to get the desired results is difficult across thousands of items and hundreds of locations. The optimization solution allowed us to identify those areas and the potential impacts behind them.

 

We saw our prices through our customers’ perspectives.

One of the most fun things to do was turn off all the pricing rules and see what the system would recommend based on our goals and customers’ behavior. Obviously, we would never implement these results, but it provided incredible clarity to the potential if you took off all your constraints. It also gave us great insights into which pricing rules we should change and whether they were beneficial or overly constraining.

 

We strengthened negotiations with vendors.

We had two vendors that sold the same commodity item. One vendor recommended keeping the price the same, but the other gave a better cost. We found no benefit for the retailer in keeping the items at the same price, so we slightly raised the price.

But that wasn’t the best part. Naturally, the vendor saw that we raised his prices and wasn’t thrilled. We sat through this well-thought-out presentation about why the two items should be priced, how consumers viewed the product, and what it meant for the retailer regarding sales and benefits. The presentation was solid.

But when it was all done, I’ll never forget what happened. The merchant, who politely sat through the entire presentation, turned to the vendor and said, “You cost me five cents more per item. My data says consumers don’t show any preferences between the two. If you want me to be the same price as your competitor, I want the same cost. Otherwise, have a good day.” She got the cost reduction.

 

We were dynamic for the market.

Before we started, the retailer rarely made price changes to realign to market conditions. On average, a price change occurred once every six months. We started with every category quarterly and then reevaluated monthly. That initially made many people nervous, but the constant refinement and category improvements made people believers. Today, pricing can be even more responsive to competitive, consumer, and market changes.

 

It enabled regional pricing.

Before the implementation, the retailer priced nationally, except for Alaska and Hawaii. Stores in those two states saw an arbitrary markup because of the extra supply chain costs. The solution allowed us to develop pricing zones and price more competitively in markets.

 

Our pricing rules were enforced.

Every retailer has pricing rules: size rules, brand rules, parity rules, margin rules, competitive rules, price points, when you can raise or lower a price, and on and on. Pricing rules can quickly overwhelm a person and result in mispriced merchandise, especially when you increase the number of price changes and zones like we did. We didn’t have to worry about that.

 

We accurately accounted for cost changes.

Before we started, one of the worst processes was updating the price when a cost change occurred. These changes weren’t always communicated or entered into the system, so prices weren’t adjusted when a cost change occurred. We implemented a rigorous process to ensure all cost changes were known, accurate, and updated. That allowed us to reevaluate prices with the correct data.

However, there were some surprises. Everyone was so conditioned that the price would increase by the expected margin when a cost change occurred. It struck us that this did not always happen. Sometimes, the recommendations did not always pass along the total increase to the price; instead, they found other opportunities to absorb the impact.

 

We became strategic.

Some merchants know their category inside and out. I’m not a merchant. I don’t have years of experience within a category. I don’t know their products, assortment, or customers like them. I didn’t have a history of managing through different periods. However, we recommended prices based on their input, strategies, and goals. And we would explain what was going on back to them. We built trust, improved communication, and uncovered insights about the category the merchants were unaware of. We became strategic advisors, whereas before, the “pricing people” had just entered the new prices.

 

It was a fast implementation.

My frame of reference was retail ERP. My most successful implementation took nine months, which was unheard of at the time. But my colleague and I still say we could have done it in six. This pricing implementation took three months. Today, it can be done even faster.

 

We had good data, and we validated it.

It helps to have good data, and it helped that I knew the retailer’s system intimately. Yes, interfaces can be tricky work. Yes, gathering two years of sales history can be rough. Yes, your back office technologies affect the process. But these are typical, day-to-day challenges of an IT person. They live and breathe this. They can do it if you prioritize it and support them.

I’m going to be fair to my fellow IT professionals. They are asked to do a lot and work with some garbage systems and data. No one might know or even care until they start looking at it. So be kind.

Regardless of the data quality, you must validate your data extractions and interfaces to ensure nothing weird happens. And when it does (because it will), take the opportunity to fix it for the betterment of your business. Often, these things get ignored until there is a reason to fix them. We did that, which was also how we identified and fixed the cost change issue I mentioned.

We had a change management plan.

We knew there would be pushback in this effort from the category managers, merchants, and store operations. Some people would challenge, others would be supportive, and the remaining would be neutral across executives and workers. This reaction is normal. Our success hinged on communicating with the executive stakeholders and the underlying merchants and operators. Everyone must be brought along. And no. Training people on the system, giving them a presentation, or telling them they must do it does not ensure change and compliance.

We had strategies for going live: How would we convince people? Who were we going to start with? How would we ensure that person saw the success? How would we ensure we were taking their feedback? How would we manage roadblocks? What were the expectations? It seems like a lot, but if I sum it up, it’s guided by a simple principle: Communicate, listen, and respect people’s experience, knowledge, and feedback.

 

What could we have done better?

Ugh, I’ve got to admit this one was tough. It’s been many years since that first implementation, so the science, features, reporting, and technology are leaps and bounds ahead of where they were. So there are some real apparent things I wish we had then, which we have now. So, that is off the table for discussion. But I did come up with a couple of thoughts.

I wish we had transitioned better. I was part of a SWAT team that came in, got it up and running, and proved the benefits. We didn’t effectively transition to other team members when we were moved off, so there was a lot of lost momentum.

We did a test and control group on the price changes. While that helped with measurement, it wasn’t the best idea because of the lost opportunity. It works for a time, or if you are trying to pilot something, but that didn’t align with our situation. There are better ways to measure price impacts that are more accurate. However, they are more complicated to explain—and, naturally, more difficult to trust.

And there is one final lesson I learned from this experience. You probably would have thought the executive management team would have been thrilled that we made $45 million in extra profit in three months. But that wasn’t the case. They planned for $60 million, and we were short. So, my last lesson is to ensure your business case is solid and can be delivered upon (under promise – over deliver). Thankfully, I did not come up with the numbers on that one.