Something we see often, especially from larger clients, is a desire to build a dynamic pricing solution in-house. The draw of keeping your pricing information in-house is obvious, and at first glance, it might seem relatively easy to do.
But what actually goes into a dynamic pricing solution? The short answer is: a lot.
If you’re thinking of building your own dynamic pricing solution in-house, we understand. But with 7 years of experience, we thought we would share some insights on what you should consider in your home-built solution.
The 4 ingredients in a dynamic pricing solution
No matter what sort of dynamic pricing system you want to build (or how complicated you want to make it), there are four main components in the process you should consider:
- The data
- The pricing logic
- The automation
- The user interface
The first two components make your dynamic pricing functional. The second two components makes dynamic pricing a success.
Data encompasses many different data points, both internal and external.
Internally, consider product purchasing price, stock levels, sales information, and more. Each of these data points help you decide how to price your products based on known information.
The most important external data source is your competitor data. You want to understand which competitor is selling the same product against what price and with what kind of delivery costs and times. This data can come from a variety of sources, such as scraping data directly from competitor websites or via comparison shopping engines.
Collecting this data sounds easier than it actually is. To have a full overview of the markets, you need:
- Connections with multiple data suppliers (the more the better), for a complete understanding of your market
- An internal data collection program for collecting purchase price, stock levels, marginal and logistical costs, product lifecycle, and more
- The knowledge to understand what that data means for your business
- The technical ability to validate data and ensure quality
Data suppliers are especially important, and the general rule of thumb is the more connections you have, the better. Sometimes one data source can’t match all GTINs and you need to supplement it with another.
Ideally you will get data from two sources: direct scraping data from competitor websites for accuracy, and comparison shopping engine data to understand where your competitors are advertising. Rates for data partners (like scrapers) can be expensive, and going with a cheaper partner will cause data quality issues. And data quality is of the highest importance when it comes to dynamic pricing.
You also need the ability to process that data multiple times per day, because some of these variables are subject to frequent changes. Competitor prices, for example, update repeatedly throughout the day, so it’s easy to fall behind the market.
Pricing rules and pricing logic
Once you have a dataset of all your competitors and internal data, it’s time to make decisions about how you want to react to changes in the market or changes in internal variables. In a dynamic pricing tool, these decisions can be made by pricing business rules or by more advanced predictive algorithms. We’ve seen that a combination of both works best. The collection of all these rules and algorithms forms your pricing strategy.
Pricing rules, at their core, tell your dynamic pricing module how to act in a given situation or when a data variable changes. Some examples of simple pricing rules are:
- Always be the lowest price on the market by 5%
- Always be the highest price on the market by 5%
- Match Competitor X’s price
In reality though, your pricing rules will get more complex than this. You can follow a group of competitors, use historical data and past performance to calculate new prices, and even include weather information to react to sudden changes in demand in the market...there are a lot of opportunities!
Your tool should use these rules to suggest “price advices” for every product in your assortment (or any product that you run through the dynamic pricing software). Your team should trust these price advices, understand how the tool calculated the price, and be willing to use those prices without a second thought.
You can build pricing rules and pricing logic, but doing so requires some imagination. You need to think about your entire strategy upfront, then ask your IT department or BI team to build rules for every situation. Once they build these rules, you’ll have a library at your disposal. It is key that a particular pricing strategy can be applied to each subset of your assortment, as different commercial strategies can apply for each part of the assortment (for example, a year round electronics category vs. a seasonal garden category).
However, building these pricing rules does take time because they are so complex, and your development team needs to build them in a way that’s also easily explainable. If these rules are not transparent, your pricing teams won’t trust the price advices.
Additionally, once you implement those rules there is one element you can’t control: how the market will react to your changes. Your system needs to be smart enough (and agile enough) to use these reactions in future price updates.
The first two steps of the process are crucial to getting a price advice that you trust and which reflects your overall commercial strategy. But for a system to truly add value, you need to automate all price changes.
The first point for automation is the data collection, specifically the competitor pricing data collection. Competitor prices change continuously and you need to be able to adapt over your full assortment without any manual steps. Manually validating price checks cost pricing teams as much as 10 hours per week per person, and the work is tedious. You should automate this part of the process to not only save time, but to also improve the overall working life for your team.
You also need to automate price updates to your online store. After your software calculates the new price advices, it can upload those prices into you e-commerce platform that displays the prices on your website, comparison shopping engines, or even electronic shelf labels. This automation step also saves crucial time and allows your store to stay agile as the frequency of market changes increases.
Automation is obviously an important part of the dynamic pricing system.
To get the most out of dynamic pricing, you should automate your entire process, from data collection to price updates. If you don’t automate the entire process, you’ll quickly fall behind the market.
But doing so removes the points for manual data verification and validation. Data validation is a crucial aspect that makes pricing automation a success. Without it, your team won’t be able to trust the recommended prices or use the insights to build more profitable strategies.
So before you begin with automation, you need to trust the system completely with your pricing data, and feel comfortable that your shop is safe. And to do that, you need to build data validation into the system, and go thorugh comprehensive data validation tests that let you evaluate the entire chain, from input to export.
Some areas to ensure safety include:
- The consistent quality of the data you import into the system. How will your dynamic pricing tool access high quality data?
- Price advice boundaries. What is the maximum or minimum price acceptable for each product?
- Failsafes. What should the system do if the chain breaks?
Building a basic infrastructure for data validation is difficult, but not impossible. But the reality is that if you want to build an agile system where pricing rules can easily be added or changed, the data validation process quickly becomes complex.
User experience and interface
Up to this point, we’ve focused on the back end of the dynamic pricing tool. But now it’s time to think about the daily use of dynamic pricing. How can you encourage the adoption of dynamic pricing and make it an integral part of your teams’ workflows?
The answer lies in the user interface of the portal you build to manage the dynamic pricing system. You can’t discount the value of user design and experience, and it might even be the biggest barrier to adoption beyond building a platform.
Your end user is your pricing and category managers. From an interface perspective, there are two goals to help them :
- Give the end user insights that make their job easier
- Allow the end user to continuously iterate on the strategy without the barrier of IT
Insights give your end users confidence in the tool, as well as the information they need to build better pricing strategies. These insights will help your team react to market trends, for example, or quickly respond when a competitor runs out of stock on a popular product.
Pricing and category managers should also be able to use the tool freely and make changes to pricing rules as needed. They should be able to do this without calling in development or your BI team every time they want to make an adjustment.
Additionally, design makes a difference. While the interface doesn’t need to be pretty, the better designed it is, the easier it will be to use. To build a proper that suits your user’s needs, you can hire consultancy agencies or do your own internal tests with your team.
IT investment for dynamic pricing
So how much time and labor does it take to build a proper dynamic pricing solution?
A conservative effort is 1-2 years for a team of developers, and it takes work from both IT and the business side of your company. This period will result in the minimum viable product for the data collection system, establishing the pricing rules, building price advice algorithm, portal, and a repricing tool (if you choose to automate your repricing). Building the infrastructure for dynamic pricing is a time-consuming and laborious task.
However, this two year period is just the start of the journey: you also need IT and development resources to maintain the tool after you’ve built the infrastructure. And for the tool to stay relevant and useful, you need to invest development time into iterating and improving the tool continuously as business demands change.
So the short story is that your dynamic pricing tool requires a significant investment from your developers in house. And as Berend van Niekerk says in Episode 5 of Price Points:
That's a lot of work. And for every company the I.T. resources are really scarce. So then it's a question — do you want to invest four or five developers full-time into building dynamic pricing system? Or do you want to invest those guys into building your e-commerce platform or anything that's important for you or for your everyday sales?
That question is one that you can only answer for yourself.
Building a dynamic pricing software is a much bigger beast than most companies expect. It requires significant investment in time, energy, and money, and is an ongoing process that you need to continually maintain and update.
Now, that doesn’t mean you can’t build your own tool in-house. And for some companies, that might be the best choice for you. But there are tradeoffs to consider — namely in the development capacity you need.
In many cases, a third-party dynamic pricing solution makes more sense economically, and it also allows your development team to focus on what really matters: your e-commerce platform.
Berend started working at Omnia Retail as a Consultant and Product Manager four years ago. He holds a MSc degree in Industrial Engineering and Management from the University of Groningen.