Category: Blog EN

Setting the correct price for a product or service is one of a retailer’s biggest, and most important, challenges in a competitive market. A price that is too low reduces revenue and, consequently, profit. Conversely, a price that is too high reduces demand and also reduces revenue. While the unit profit on sales is higher, it may not offset the losses associated with reduced volume.

Determining the optimal product price – where to start?

The difficulty in determining the optimal price of goods is due to the number and complexity of factors that affect it. These include the prices of complementary and substitute products. To the list can also be added promotions, advertising, the activity and offerings of competitors, the economic situation of customers, their tastes and preferences, the cost of purchasing goods, logistics issues. Additional challenges are associated with products whose demand is seasonal and is due to weather conditions, for example. A classic example of such an assortment is ice cream, the demand for which is closely related to the air temperature. Other such products that can be mentioned are sunglasses, skis, cool drinks, winter jackets, swimwear, hotel services, tourist flights.

The paradox of price

There is often a price paradox with these types of products. The classical economic theory of supply and demand says that the higher the price, the lower the demand for a good or service. This is usually the case. However, in the case of eminently seasonal products, this relationship can be shaken. After all, it often happens that the highest prices are quoted during the peak season for a given product. That is, when sales are highest as well. Sellers, knowing that it is during this period that consumers most need the product and are most likely to buy, take advantage of this by raising prices. A situation of this kind is illustrated in the chart below. You can see a gentle but clear positive relationship between price (horizontal axis) and volume sold (vertical axis).

Just looking at this graph, which is essentially a simple price-demand model, one could draw the naive conclusion that raising prices increases sales. As we know, this is not true (except for a very narrow and specific group of luxury goods). By raising the price of ice cream between October and March, we will not succeed in increasing its sales.

The relationship observed in the graph is the result of the intertwining of two factors. First is the effect of price and air temperature on sales, and second is the effect of temperature (season) on price. The latter is precisely related to the seller’s decisions to adjust the price to the increased demand.

Is there an ideal price estimation tool?

Thus, solving the puzzle of the real effect of price on sales requires a somewhat more complex approach than the analysis of the correlation between price and demand (an example of which was illustrated in the previous chart). Randomized experiments (e.g., A/B tests) are the ideal tool for estimating such effects. In theory, one could imagine that a retailer would randomly change prices so as to test different variations and combinations – resulting, for example, in an increase in the price of ice cream in December, or its drastic reduction in an exceptionally hot June. In practice, however, this is hardly feasible, and if only on a small scale and for a limited time. This is because it is a very expensive experiment.

Selling products at suboptimal prices drains revenue. In addition, frequent and unpredictable price changes can negatively affect consumers’ experience and induce them to switch to competitors. A practical solution, then, is to use the data we already have that are not from the experiment to estimate the effect of the factor of interest (in this case, price) on the outcome that is important to us (in this case, sales). It is possible, although it should be noted that this is not a trivial task and requires an elaborate mathematical apparatus. In today’s article, however, I will not focus on the mathematical, statistical and philosophical nuances involved in causality analysis. Instead, I will show the possible results of this kind of analysis and their practical effects.

Seasonality vs. product price

The solution requires at the outset the imposition of a “model” on the data that reflects our understanding of how the system we want to analyze works. This requires both common sense and expert knowledge. We can describe our assumptions using a graph, as illustrated below.

First of all, as you can see, we set assumptions as to the direction of influence of the various variables. The price changes under the influence of temperature (or, more precisely, by the actions of the retailer resulting from his knowledge of the effect of temperature on consumer behavior). The retailer, on the other hand, by changing the price, is not able to change the air temperature – I think this is not a controversial assumption – hence the arrow points in only one direction. Temperature also has a direct effect on sales (when it’s warm more people are eager for ice cream). In addition, we take into account that other factors that are beyond our observation may also affect the price and affect sales. This is, of course, a very simplified model and can easily be expanded to include other factors that affect price or sales.

Estimation of the actual effect of price on sales

Based on such a formulated model and historical data on price, volume and temperature, using advanced analytical methods, it is possible to estimate the actual effect of price on sales. In other words, we can estimate to what extent a change in price is the cause of a change in sales volume. From there, it’s only a step to putting this knowledge into practice and optimizing the price.

The following example shows the process of finding the optimal price for ice cream in the month of September with a forecasted average daily temperature of 14.2 degrees C. We can generate similar charts for any temperature value, which is important in this case. This is because the optimal price will be different when September is exceptionally warm, and different when there are records of cold weather.

September is a transitional month between the peak season and the autumn-winter season. So far, the retailer has traditionally kept product prices still quite high in September. On the horizontal axis of the graph we show the change in price relative to the base price – this is indicated by the 0 line near the center of the graph. The blue curve shows how sales revenue changes depending on the price adopted. Moving to the right of line 0 (i.e., increasing the price), we observe a decrease in revenue. A higher price negatively affects demand, and higher unit revenue does not compensate for the decrease in will. Moving to the left of the 0 line, we observe an increase in revenue, although only up to a certain point. Beyond it, increased sales volume ceases to compensate for the drop in price. This point marks the point of optimal price – indicated by the vertical dashed line. The model suggests that the optimal price is 80 cents lower than the existing base price.

Following the recommendation and lowering the price results in an increase in turnover of just over 26% compared to the base scenario. This is illustrated in the chart below.

The method presented in the article offers tremendous opportunities, using the latest research advances in artificial intelligence and causal analysis. In practice, of course, for greater accuracy, the model should also take into account additional factors such as prices of other products, prices at competitors, advertisements, flyers, newspapers, offers, promotions, macroeconomic factors. The full solution also gives much more in-depth insights. Simulating the impact of pricing decisions on sales volume and revenue helps you make better decisions. These, in turn, translate into measurable financial results and can give a significant competitive advantage.

Acquiring a new customer is more expensive than keeping an existing one

This is not just an oft-repeated marketing truism. Research cited in the Harvard Business Review shows that the cost of acquiring a new customer can be 5 to as much as 25 times the cost of retaining a customer, depending on the industry. And improving retention rates by just 5% can translate into as much as a 25% increase in profits. So how do we combat customer loss and increase retention? How can data analytics help us do so?

Customer churn (churn or attrition) is an inevitable phenomenon and it is impossible to eliminate it completely. Some customers, regardless of the measures taken against them, leave. For example, because they move out of the company’s area of operation or cease to be a target group and no longer need our product. The remainder, however, give up, opting for a competitor’s offer. These departures could have been prevented. If action had been taken. The right actions, at the right time. The keys are:

• Predicting the risk of customer departure with sufficient accuracy and in advance
• Understanding the factors that influence the risk of customer loss

The solution to both problems can be an anti-churn predictive model built using machine learning. Such a model is capable of predicting the risk of losing a particular customer. In doing so, it identifies the most important factors associated with an increase in this risk both generally for the entire customer base and individually for a single customer in his or her specific situation. Such predictive models can use any definition of “churn” and are applicable both to businesses where the departure of a customer is clearly marked in time (e.g., expiration/termination of a contract) and those where the customer simply stops returning and making further purchases.

The most important factors determining customer departure

As we mentioned, the predictive model helps identify the most important factors influencing the risk of customer churn. The charts below are from the actual predictive model built on one of Data Science Logic’s contractors. Only some of the variable names (including product category names) have been changed. It is worth noting that this is an industry characterized by a relatively low frequency of purchases (a few times a year on average) and high customer turnover.

The chart at the top shows the customer characteristics that most explain the likelihood of leaving. As you can see, the key variable is the number of days since the last visit with a purchase. This is not surprising. The longer a customer has been gone, the less likely they are to return. However, the model allows you to pinpoint when the increase in risk is greatest and when you need to take decisive action. As you can see in the bottom graph, up to about 365 days the risk increases linearly. After more than one year of inactivity, the risk curve becomes steeper. This is the last moment to undertake an anti-churn campaign.

Also of interest is the second most important variable – the number of visits with a purchase of an “A” category product in the last 12 months. These products are exceptionally well regarded by customers and have a positive impact on customer satisfaction and retention.

In addition to general conclusions about the factors influencing the risk of losing customers, the model allows us to predict the probability of losing a particular person and to identify the specific characteristics that, in his case, increase or decrease this risk, as shown in the chart below. In his case, the risk is relatively low (35.5% compared to the baseline 49.6%). The risk is reduced by, among other things, the average value of the visit and the number of visits over the past year. However, the customer does not use the products of the aforementioned “A” category, which increases the risk of leaving. Encouraging them (e.g., through an appropriate campaign) to try products in this category would likely lower their risk of leaving even more.

Dealing with customer migration is one of the most important challenges facing companies today, given how expensive it can be to acquire a new customer later on. With antichurn modeling, we will learn which customers are likely to leave and why, the signs of increasing risk of leaving, and how best to prevent them from leaving.

Up to 80% of promotional campaigns do not bring any noticeable increase in sales or the cost of granted discounts is higher than the margin generated on additional turnover – proves a Boston Consulting Group study. The former can be abandoned without fear of lost sales. It is necessary to resign from the latter ones, because their financial effect is negative. Meanwhile, the strategy of many retailers is based on promotions. This applies to most industries. And what is more, it concerns both stationary channels and online trade.  The pressure of competition, the fight for market share and consumer habits mean that the annual number of offers often runs into hundreds or even thousands. So the space for optimization is huge.  But how to distinguish profitable promotions from those generating losses? How to measure the real effect of a promotion? Is it possible to predict the outcome of a promotion before it even starts? How can we better plan promotions?

Measuring the effect of a promotion

The key to success is to measure the effects. Without knowledge of the actual results of a promotion, managers are doomed to strategies such as copying the “proven” calendar of previous years, reacting to the actions of competitors or simply their own intuition. In theory, the task may seem simple. It is enough to compare sales during the promotional period with… Well, exactly… with what? Various possibilities come to mind: maybe sales before and after the promotion period? Maybe sales in the same period of previous years? How about excluding certain outlets from the promotion and using them as a comparison group?

However, demand for products is often seasonal and changes regardless of promotions. The previous year may have been different due to the macroeconomic situation or the entry of a new competitor into the market. Selecting a representative comparison group can be difficult, if not impossible, for online retail. Additionally, promotions overlap. Some are heavily advertised in the media, others only at the store shelf. The real business world is complex and overly simplistic methods of analysis can lead to incorrect conclusions. 

With help come advanced methods of data science. Thanks to them, it is possible to develop a model that takes into account many factors influencing the effect of a promotion. From the assortment covered, through the amount of discount, additional conditions and mechanics, other concurrent campaigns, to the weather, advertising activity and actions of competitors. It is necessary to collect accurate data on historical promotions and their characteristics. A properly prepared model allows to isolate the influence of individual factors on sales. This enables understanding their impact, both for historical campaigns and predicting the effect of promotions yet to be planned. 

The graph below shows the achievable prediction accuracy of the model. The blue line represents actual sales, while the red line represents sales predicted by the model. As you can see, both lines are very close to each other. In practice, it is impossible to avoid some deviations, especially visible on days with extremely high or low sales, but the model identifies trends and directions of changes very well. 

Testing a promotion before it begins

A prepared model with adequate predictive capability allows to forecast sales depending on the date, duration, scope and nature of the promotion. This makes it possible to simulate various action scenarios and find answers to questions such as:

– will the promotion generate additional sales?

– what is the best period to conduct a promotion?

– what is the optimum duration of a promotion?

– what kind of communication support is worth providing for the promotion?

– Is it worth running the promotion, taking into consideration all effects?

The chart below shows a comparison of sales in the scenario assuming running the promotion (blue line) in comparison with the baseline scenario in which the promotion would be abandoned during the period under consideration (red line). It is clear that the blue line is above the red line for most of the duration of the promotion. Particularly large increases are seen at the end of the promotion period, as well as at the beginning. The timing of the promotion was clearly communicated to consumers in this case hence the accumulation of sales. In the scenario without the promotion, sales would be more evenly distributed with only a periodic weekly cycle visible. The average sales level, as the graph shows, would be lower. So it seems that promotion has a positive effect on sales and should be profitable. But is it really?

A comprehensive model of promotion

When evaluating the effects of a promotion, one should look not only at the sales increase, but also at other phenomena related to the promotion that would not occur if there was no promotion. It is primarily about the shift of sales in time (in the example visible in the period just before and just after promotion), as well as about the impact of promotion of a given assortment on other product categories. For complementary categories, we expect a positive impact on sales. However, for other categories (for example, non-promoted substitute products), the effect may be negative. It is the interaction between these different effects that determines the overall profitability of the promotion. It is therefore necessary to estimate them precisely. A model built and tested with the use of data science methods makes it possible.

The diagram below visualizes the decomposition – the breakdown of the total effect of a promotion into individual components. The base sale is the sale that would have been realized if there had been no promotion.  The additional promo sales were estimated at 5.3 million using the model. This is how much more we sold of the products promoted by running the promotion.

Complementary category growth is the positive impact of the promotion on categories typically co-purchased with the promoted products. Cannibalization is the negative influence of a promotion on sales of other product categories – in this example it amounts to 9.7 million, thus cancelling out the entire positive effect of promotional activities. Additionally, effects related to the time shift in sales – before the campaign (the waiting effect) and after the campaign (the effect of buying on stock at lower prices) worsen the result of the action by another 1.3 million.  After taking into account all the above-mentioned effects of the promotion, total sales were 6.8 million, i.e. 1.8 million less than in the scenario without the promotion. Therefore, the real incremental impact of the promotion on sales is negative, i.e. the promotion was not profitable.

Summary

In conclusion, a simplistic approach to promotion analysis and limiting itself to the effect of increased sales of promoted products during the promotion period may lead to incorrect conclusions and suboptimal decisions. Only a complex analysis based on a wide range of data and using advanced data science methods can answer the key questions from the point of view of promotion planning. Precise promotion models allow to accurately estimate particular effects, to accurately simulate alternative scenarios and to optimize not only single promotional campaigns, but the whole promotional calendar.

The stationary sales in the vast majority of industries plays a key role. Despite the observed dynamic growth of e-commerce, this will not change in the coming years. The opening of a new point of sale usually entails a significant investment related to the construction or rental and adaptation of premises, recruitment of employees as well as changes in logistics. Additionally, the potential negative impact of a new outlet on the existing ones is significant. Therefore, decisions to expand the sales network are associated with high risk. In today’s article, we will show how data science combined with geospatial data can help mitigate these risks and facilitate better decisions.

Key questions

In the context of point-of-sale locations, questions that data science can help answer include:

1) Is this a good place to open a new store? 

2) Will the new store not “cannibalize” the sales of my existing stores?

3) How many stores should I open, where should I open them, what should the optimal network look like?

4)Which stores should I close? What will be the net effect of closing a store?

5) Is the existing store using the potential of its location?

6) If I don’t open a store in a particular location but a competitor does, will my existing stores be negatively affected? Which ones? How much?

Today I would like to focus on the first two questions and show you how data analysis can help you make the right decisions.

Data, data, data…

To begin with, it is worth taking a moment to look at the sources of the data used in the analysis. These can be divided into internal data and externally necessary data. Key internal data includes:

– historical sales data, 

– outlet characteristics (space, nature of location – gallery, stand-alone, etc., range of assortment available)

– local activity (promotions, media presence, leaflets, newspapers, billboards),

– address data of points.

The data that need to be obtained from outside are mainly:

– data on population, demographic characteristics (age group distribution, gender), income and purchasing power,

– Data about the road network, its quality/class and traffic volume,

– geolocation of competition points,

– travel time to own and competitors’ outlets by different modes of transport (depending on the nature and density of the sales network, different modes of transport may be relevant).

Some data may be available only at the level of the whole municipality (especially data from the Central Statistical Office), but where possible data of the highest granularity should be used. There are sources from which data can be obtained for individual address points (specific blocks). 

When analyzing and presenting data, a reasonable compromise between detail and total may be a so-called kilometer grid. The map is then divided into squares of 1km side length. Examples of such maps will appear later in this article.

Why is accurate geographic data important?

Below is a simple example of the differences in conclusions that can be reached depending on the data available. The map on the left shows the distance from the store (up to 20 km). This is a very simple measure to calculate. One might think that it would be a sufficient approximation of the time to get to the store. Unfortunately, as you can see in the map on the right, taking into account the distance from the store alone is misleading. Only showing the actual travel time on the map shows a realistic picture of the store’s range. It can be seen that the store’s range extends along traffic routes (in this example, radially spread), and areas that are close to each other in reality may have different travel times. Over-simplification and abandonment of accurate geographic data leads to incorrect estimation of store potential and potentially wrong decisions.

In which direction is it profitable to develop a chain?

We will now analyze the example of a chain currently consisting of 4 stores. On the map below you can see their range. From each area (square) the travel time to the nearest store has been calculated. The management is considering various scenarios for further development. One of them is to fill the “white spots” in the network coverage. Such a move could be interesting for at least two reasons. First, there is a town in the area with what appears to be demographic potential where a new outlet could be located. Second, a new store created between existing stores could fit perfectly into the existing logistics chain.

To base the decision on data, an estimation of the new store’s potential is made and its impact on the existing network is simulated.

The map on the left shows the range of stores before expansion. Areas were assigned to the store with the shortest travel time. The map on the right illustrates how the coverage of existing locations will change after the network expansion and what the coverage of the new outlet will be. It can be clearly seen that the overall network coverage will be expanded to include new areas. You can also see that the areas of all but one of the existing stores will be slightly depleted. However, a visual assessment and map analysis is not enough to make a decision. Precise forecasts are needed. Only accurate numbers will allow to estimate the profitability of the considered investment.

Predictive model

With help comes a predictive model built on machine learning. Using a wide range of available data (sales, demographics, geography), the model allows for accurate estimation of a new store’s potential and its impact on existing outlets. The graph below shows the modeling results. The bar on the left (‘Existing chain’) represents the baseline, i.e. the projected sales level of the entire chain if the new store had not been launched. The next bar is the sales estimate for the new outlet. The result shows that it will increase the potential of the chain. However, compared to others, its contribution will be relatively lower. The new outlet will increase the chain’s turnover by about 12%. The next bars show the cannibalization of sales in the existing points. As could be guessed from the map analysis, 3 out of 4 stores will be affected by cannibalization. It may seem that no store will suffer significantly – on average only by about 6% of turnover. However, it will account for as much as 54% of the new outlet’s sales. Thus, most of the new store’s turnover would be realized at the expense of the existing stores, and the incremental impact of the new point on the chain’s total turnover would be only about 5%.

The final decision about the profitability of investment in opening a store in the considered location requires comparing incremental turnover (and margin) with the necessary expenditures and operating costs. The analysis should also include a margin forecast, as it may turn out that the new store will differ from the existing ones in terms of a typical basket of products and, consequently, their margins. It is certainly worth considering other potential locations, as the return on investment there could turn out to be higher. Additionally, the possible actions of potential competitors should also be taken into consideration. The most appropriate course of action would be to conduct a comprehensive analysis and simulation covering many potential locations.

Modern optimization methods, which we use on a daily basis in Data Science Logic projects, allow us to simulate many parallel scenarios and find the optimal shape of the network. Thanks to this, they are able to indicate which locations are worth opening and which should be closed. The final decisions always belong to people, but precise data combined with appropriate methods of analysis can help to make them.

Want to know more about customer traffic in stationary stores? Read the details of the hourly traffic forecast project and see how it will help increase sales.