Profit growth with dynamic pricing:
Benchmarks by retail industries. Factors causing slowing down and accelerating growth

How much will the gross profit of our stores increase?
This is a question that interests most retailers who are implementing dynamic pricing.

In this article, you will find:

Benchmarks for profit and sales growth in online and offline retail.
Factors that affect the magnitude and speed of growth.

This will help you evaluate the potential impact of implementing dynamic pricing in your retail segment, taking into account the specifics of your stores.

Factor 1. Retailer's specialization

Profit and sales growth varies depending on the specialization of the retail chain — such as food retail, pharmacy / drugstore, DIY, and so on.
Below are the growth ranges for some segments after implementing dynamic pricing:
Gross profit lift
Gross profit lift
Gross profit lift
Gross profit lift
Gross profit lift
Revenue lift
Gross profit lift
Drogerie, cosmetics,
Household, DIY,
Household, DIY,
Gross profit lift
Drogerie, cosmetics,
Gross profit lift
Revenue lift
Auto parts,
Gross profit lift
Needlework & craft,
Imprice customer statistics

Factor 2. Sales channel.

  • Marketplaces (such as Amazon, eBay, Craigslist, and so on)
Each marketplace has its own specifics: rules, technical features, and a composition of sellers (retailers who sell the product).

As a result, the profit growth with dynamic pricing for the same retailer varies depending on the specific marketplace.
  • Online and offline channels
When implementing dynamic pricing, sales and profits grow faster for online stores than for offline retail. This is the case in most instances; the difference between online and offline metrics can be studied HERE.

Why sales and profits grow slower in offline retail:

Reason 1 - Manual price tag changing.

The assortment of a hypermarket consists of around 20,000 SKUs; a supermarket has around 5,000 SKUs.
The resources of stores allow for changing 200-500 price tags per day, which is 1-10% of the assortment. Note that first they always change the price tags of the products whose cost price has changed and the price tags of KVIs.

A significant share of the store's profits potential is "hidden" in products from the Foreground and Long Tail segments. AI algorithms calculate the optimal set of prices for these goods:
-- If the price of a product is lowered, the algorithms increase it to the optimal value; sales remain at the same level, while gross profit grows.
-- If the price of a product is overestimated, the algorithms reduce it to the optimal value, and gross profit increases due to the growth of sales.
The optimal price set ensures the growth of sales and profits, recovering the reduction in margin on KVIs. To calculate the optimal price set, the AI algorithms require sales data at different price levels.
At the same time, the "Exploration or Exploitation" dilemma is solved: the task is to spend as little time as possible searching for the best price and to exploit such a price for as long as possible.
On average, the algorithms need to test 3-4 prices at the start of optimization. In offline retail, for relatively rarely purchased Foreground and Long Tail products, each price should be unchanged for 2-3 weeks to measure the demand at that price.

Foreground and Long Tail goods are thousands or even tens of thousands of SKUs with relatively rare sales. The store changes 50-150 price tags for these products per day; and each price should be exposed for 2-3 weeks.

● Of course, price optimization goes much faster online. An online store can change all prices daily, and the price efficiency level can be calculated in a matter of days, not weeks.
Reason 2 - Lack of price change history.

The process of changing prices online is simpler than in brick-and-mortar stores. That's why online stores often accumulate significant sales history at different price points. This data allows reducing the number of pricing experiments, therefore algorithms optimize prices faster.

Offline retailers often lack price change history for non-food items, so training algorithms requires more experiments, and price optimization takes longer.

Reason 3 - Offline stores' channels for notifying customers of prices are less efficient.

Offline stores have fewer communication channels with customers. Customers' reactions to prices take more time (they need to get to the store), and there are geographical limitations. In grocery retail, many customers only learn about a new price during their visit to the store.

Online stores have very high price transparency. Customers can open five websites in a couple of minutes and compare prices. An attractive offer generates traffic and sales much faster than in offline stores.

Reason 4 - Material barriers in the brick-and-mortar store.

At any price, a store can't sell a product if:
It was in the store's warehouse, but hasn't hit the sales floor.
It's out of stock in the store's warehouse.
It doesn't have a price tag.
It's not visible on the shelf.
And so on.
You can learn about the most common "interferences" and "noise" in offline stores HERE.
Eliminating such "material barriers" increases store sales and profits. This is a very common indirect effect of implementing dynamic pricing.

The direct effect of implementing dynamic pricing in brick-and-mortar stores is seen later, after the "barriers" problems have already been resolved.

There are far fewer such delays due to "interferences" in online stores. Errors do occur, but they are much easier to identify and eliminate.

➜ Outcome:
In retail chains, the main increase in profits often occurs after the pilot project has ended.
An example of a delayed effect in food retail:
Dynamics of LFL in pilot and control stores.
At the start of the rollout, the pilot store outperforms the control store in revenue by 1.1%.
After six months, it's up to 3.6%. After a year, it's up to 8.6%.
To enlarge the image, click on it.


Factor 3. Retailer's IT infrastructure errors.

In some retail companies, internal price-setting processes are working incorrectly due to ERP and related systems settings and suboptimal logic embedded in them.
Examples of IT errors from practice in 2022-2023:

Example 1.
Some stores occasionally fail to "receive" new prices from their ERP due to net connectivity issues, which are not recorded anywhere. Sales are made at outdated prices, and it reduces profits.

Example 2.
When a new batch of goods with a new cost price arrives, the new price reaches the store with a one-day delay. An IT specialist built this business logic into the retailer's IT system; but the commercial department is unaware about that.
Outcome: on the day of receipt of goods, sales are made at the old price. Sometimes, the "old price" is lower than the current cost price; then the store sells "at a loss" all day long.
Before launching dynamic pricing, retailers often do not aware about the weaknesses of their IT system related to price setting.
Dynamic pricing requires monitoring of sales results and prices. Dedicated analyst sees odd sales fluctuations, makes questions to the store, and detects a lot of internal errors. For example, the process of taking SKUs out of promotion is not formalized, there are accounting issues, the repricing logic generates losses in some cases, etcetera.

If a company has IT errors related to price data exchange, profit growth will slow down; it will require to first eliminate the discovered "bugs" in the IT system and processes. The optimal price will not be efficient if customers do not see it.
Such errors can result in 1-3% loss of gross profit.

Positive moment:
After the IT errors are eliminated, there will be a 1-3% increase in gross profit.
Often, before launching dynamic pricing, retailers do not have tools to track such errors.

A positive point:
Most technical errors can be detected using the "List of SKUs with low or negative margin" report.

Factor 4. Level of pricing automation before dynamic pricing implementation.

The more manual actions were in the retailer's pricing process before the implementation of dynamic pricing, the higher profit the retailer will gain after the implementation of dynamic pricing. In such cases, the "low base effect" works.
An example of growth after the launch of dynamic pricing in a grocery retail chain with an initial low level of pricing automation. The dark purple line is the difference in year-over-year growth in gross profit between the pilot and control store.
To enlarge the image, click on it.


Factor 5. Retailer's Goals.

Not all retailers aim to increase profit with pricing automation.

For some retail companies, the main goal of automation is to relieve talented employees.
Such retailers want their commercial department to work more efficiently with product matrix, assortments and suppliers, to search for new demanded positions, strengthen supply chains and stop spending 30-50% of working time on routine price calculations in Excel.

In projects with such goals, profit growth is usually lower than in other retailers.
The reason: responsible employees focus on other aspects and do not seek additional growth opportunities. It is enough for them that the platform generates correct prices, and profit growth covers the cost of implementation.

Factor 6. The role of analytics in the company's strategy and willingness for experiments.

If the company's work culture is focused on constant analysis, searching for growth points, and testing hypotheses, the result of dynamic pricing implementation is higher. Dynamic pricing software makes conducting price experiments and introducing new pricing strategies and approaches dramatically easier and faster.


Factor 7. Planning of the pilot; choosing test trade points and test categories.

The size and speed of the increase in gross profit within a pilot test depend on several parameters of that pilot.

One of these parameters is which product segments and categories are involved in the pilot:

If only categories with small revenue and profit and with rare sales are introduced in the pilot, the increase in money terms will also be small.

If a wide range of products are included in the pilot, including popular items with high sales, the increase in money will be significant and will occur faster.
To conduct a dynamic pricing pilot, experts chose a product category with a revenue of 1,000 euros per month and a profit of 150 euros.

After price optimization, revenue increased to 1100 euros per month, and profit to 250 euros.

The percentage growth was high, but the increase in money terms was insignificant. The implementation of dynamic pricing during this period will not recover costs of automation. It will happen later, when a wider range and/or a larger number of stores are connected to dynamic pricing.
Recommendations for pilot planning can be found HERE.
Talk to Imprice pricing experts: