Do you also have 50% of marketing shares playing a minus on sales?
Imagine that you are producing juice. You deliver this juice to large retail chains and put it on a shelf next to a dozen more similar ones. You want to sell it more, and decide to do a share for this. For instance:
How to make a decision?
Now another example. You have a ton of yogurt in your store that expires in a month. When and what actions should be done? When to reduce the price, when to include a bonus to sellers, when just to merge them at cost?
There are expert systems for making such decisions. I’ll tell you about them.
In the general case, the task boils down to three subtasks:
The most common example, of course, is evaluating the effectiveness of a marketing campaign before it starts. You definitely need to know how much money and how you will earn, how much goods will go, and so on. This can be considered hands, or it is possible with the help of expert systems to solve just such problems.
We have our own special atmosphere in Russia. Therefore, quite often, such a solution is needed both by suppliers of goods (they hold promotions in stores), and by retailers themselves. It is very difficult to evaluate the direct economic effect of marketing actions in practice, and in the context of a limited budget, nobody wants to spray money just like that. But there is a tool that allows you to evaluate the effect of any marketing campaign, to predict this effect (based on several factors), to say - is it worth it or not, and so on. And as a result, we at least increase the effect of our stocks, as a maximum - we greatly save on cutting off unprofitable ones.
What does it look like?First, data on sales and stock parameters is loaded into the system. It is necessary to determine what marketing campaigns were conducted, what was the lift effect on sales during the campaigns. Next, a fairly simple data analysis will take place, and the system will display a list of parameters for new stocks. With recommendations.
Why do you need to do this automatically? Because in reality, you can never count only one channel. For example, you could have advertisements on two TV channels, on three radio stations, posters were hanging on points, plus flyers were distributed near the metro, and once you still gave a 10% discount on goods. To understand which of these channels to use next time without any tool is almost impossible, but you need to make decisions about the next and next promotions.
There is a history of basic sales in a particular region. Prior to that, you put promo stands there (special bright shelves) in the trading floors, for which you paid the retailer, but in different places of the shops, for example, at the entrance, next to other chocolates or near the cash desks. You upload data, for example, in Oracle Demantra and calculate how sales will go in one case or another. And you see that they remained in advantage only in individual cases, for example, in hypermarkets when the stands were next to the other chocolates, and in small supermarkets - near the cash desks.
Now you want to put up such a promotional stand in several more stores in this region, but you don’t know exactly where it will work and where it will not. Again, look at Demantra and, based on the results of the analysis there, make a decision.
In one network, you tried promotional stands in some stores, in others priority placement, in the third you started sampling - tasting stands in trading floors, in the fourth you made a discount, in the fifth you set a bonus for staff, and in the last two stores you tried to load sausage in a new bright packaging. Moreover, everything was done in one region, for example, in the Volgograd region.
And now you are being put into a new network in the same region, and it would be great for you to learn how to distribute the budget as efficiently as possible. Which channels are best used, how much to spend on each of them, how long will each be effective? The system can use the accumulated statistics of various promotions and offer the best options for events for a new (similar in format of stores, type of customers, etc.) retailer.
It is clear that at the same time, you have factors parallel to the forecast of sales - seasonality, activity of competitors, availability of goods at points, and so on. The marketing campaign management system allows you to isolate all these factors and calculate the unattainable - the effect of marketing campaigns.
More information about individual modules has already been one , two and three topics.
More than once I have come across the fact that those who understand how to count - try to count in one way or another themselves (at least on paper, at least in XLS). And from a certain volume, they begin to understand the need for an expert system. But there are other manufacturers where either marketing is not so strong, or there are a lot of involved promotion channels. Often in such difficult situations they put pressure on an already tested channel (usually one, and to the maximum), and then they begin to shaman. In fact - they are wasting money. It is very important to understand that if marketers do not have a clear feedback, in the form of profitability of marketing campaigns, it is extremely difficult for them to build optimal promotion channels in terms of ROI. But now there are more and more promotion channels and the budget may simply not be enough for everything. As a result, half of the marketing budget goes nowhere, and the most unpleasant thing is that you don’t know which one. The systems I am talking about here help to understand very precisely which action leads to profit and which does not.
There was an example with one company producing famous clothes and shoes. There was no history of promotions. When we arrived, the first time was spent on picking up correspondence in the mail on the topic of planned promotions, comparing them with real data on sales history and product availability, finding old advertising materials, understanding what and when they were giving discounts, where they were advertising and so on. After that, we formalized the parameters of each stock, removed emissions and began to turn the chaos of unstructured data into a system. As a result, when we finished everything, it turned out that half of the stock calendar could simply not be spent, and sometimes it even turned out that the company played a discount for itself without a tangible elevator. That is, the goods would be sold the same way, at a discount, that without.
The management, seeing our work, began not only to evaluate the effectiveness of each stock, but to set the task for the departments, for example, to increase sales by 50%, and we considered the assembly of methods that would allow us to do this, since there were opportunities for combining stocks.
When you know exactly why this is necessary. For example, they realized that last year 50-80% of marketing campaigns went with an unsatisfactory ROI. Or if you want to precisely and specifically justify the marketing budget not by personal feelings, but by calculations based on the analysis and extrapolation of history.
If you think with your hands and everything works, don’t touch it. But when there will be too many factors for accounting, the volumes will be of an all-Russian scale and it will be necessary to remove the influence of the human factor - go to the systems for forecasting demand and calculating the effectiveness of marketing campaigns. One of the most difficult moments is the smoothness of the transition. There are a lot of pitfalls, and all are different. Therefore, if there are questions specifically for you, ask in the comments or write to me at AlIsaev@croc.ru, I will try to help.
- Reduce the price by 25%. It is clear that profit from a unit of juice will fall, but will it be possible to sell it so much that overall profit will grow?
- Or maybe make “buy three - the fourth for free"? Technically, it's almost the same, but in money it can turn out quite differently.
- Or maybe it’s better to pay the same money to the network so that they put it right in the center of the shelf?
- No, let's hire promo girls in miniskirts, right?
How to make a decision?
Now another example. You have a ton of yogurt in your store that expires in a month. When and what actions should be done? When to reduce the price, when to include a bonus to sellers, when just to merge them at cost?
There are expert systems for making such decisions. I’ll tell you about them.
In the general case, the task boils down to three subtasks:
- Identify all possible channels and ways of conducting the action.
- Collect data on the effectiveness of each or impose a profile of similar stocks from similar products.
- Calculate a set of measures that will give the greatest profit (or solve another problem, such as the optimal sale of goods before the expiration date).
The most common example, of course, is evaluating the effectiveness of a marketing campaign before it starts. You definitely need to know how much money and how you will earn, how much goods will go, and so on. This can be considered hands, or it is possible with the help of expert systems to solve just such problems.
We have our own special atmosphere in Russia. Therefore, quite often, such a solution is needed both by suppliers of goods (they hold promotions in stores), and by retailers themselves. It is very difficult to evaluate the direct economic effect of marketing actions in practice, and in the context of a limited budget, nobody wants to spray money just like that. But there is a tool that allows you to evaluate the effect of any marketing campaign, to predict this effect (based on several factors), to say - is it worth it or not, and so on. And as a result, we at least increase the effect of our stocks, as a maximum - we greatly save on cutting off unprofitable ones.
What does it look like?First, data on sales and stock parameters is loaded into the system. It is necessary to determine what marketing campaigns were conducted, what was the lift effect on sales during the campaigns. Next, a fairly simple data analysis will take place, and the system will display a list of parameters for new stocks. With recommendations.
Why do you need to do this automatically? Because in reality, you can never count only one channel. For example, you could have advertisements on two TV channels, on three radio stations, posters were hanging on points, plus flyers were distributed near the metro, and once you still gave a 10% discount on goods. To understand which of these channels to use next time without any tool is almost impossible, but you need to make decisions about the next and next promotions.
Now you sell chocolate
There is a history of basic sales in a particular region. Prior to that, you put promo stands there (special bright shelves) in the trading floors, for which you paid the retailer, but in different places of the shops, for example, at the entrance, next to other chocolates or near the cash desks. You upload data, for example, in Oracle Demantra and calculate how sales will go in one case or another. And you see that they remained in advantage only in individual cases, for example, in hypermarkets when the stands were next to the other chocolates, and in small supermarkets - near the cash desks.
Now you want to put up such a promotional stand in several more stores in this region, but you don’t know exactly where it will work and where it will not. Again, look at Demantra and, based on the results of the analysis there, make a decision.
Now you trade in sausage
In one network, you tried promotional stands in some stores, in others priority placement, in the third you started sampling - tasting stands in trading floors, in the fourth you made a discount, in the fifth you set a bonus for staff, and in the last two stores you tried to load sausage in a new bright packaging. Moreover, everything was done in one region, for example, in the Volgograd region.
And now you are being put into a new network in the same region, and it would be great for you to learn how to distribute the budget as efficiently as possible. Which channels are best used, how much to spend on each of them, how long will each be effective? The system can use the accumulated statistics of various promotions and offer the best options for events for a new (similar in format of stores, type of customers, etc.) retailer.
Other factors
It is clear that at the same time, you have factors parallel to the forecast of sales - seasonality, activity of competitors, availability of goods at points, and so on. The marketing campaign management system allows you to isolate all these factors and calculate the unattainable - the effect of marketing campaigns.
More information about individual modules has already been one , two and three topics.
How is the distribution of the budget in reality?
More than once I have come across the fact that those who understand how to count - try to count in one way or another themselves (at least on paper, at least in XLS). And from a certain volume, they begin to understand the need for an expert system. But there are other manufacturers where either marketing is not so strong, or there are a lot of involved promotion channels. Often in such difficult situations they put pressure on an already tested channel (usually one, and to the maximum), and then they begin to shaman. In fact - they are wasting money. It is very important to understand that if marketers do not have a clear feedback, in the form of profitability of marketing campaigns, it is extremely difficult for them to build optimal promotion channels in terms of ROI. But now there are more and more promotion channels and the budget may simply not be enough for everything. As a result, half of the marketing budget goes nowhere, and the most unpleasant thing is that you don’t know which one. The systems I am talking about here help to understand very precisely which action leads to profit and which does not.
What to do when there is no data?
There was an example with one company producing famous clothes and shoes. There was no history of promotions. When we arrived, the first time was spent on picking up correspondence in the mail on the topic of planned promotions, comparing them with real data on sales history and product availability, finding old advertising materials, understanding what and when they were giving discounts, where they were advertising and so on. After that, we formalized the parameters of each stock, removed emissions and began to turn the chaos of unstructured data into a system. As a result, when we finished everything, it turned out that half of the stock calendar could simply not be spent, and sometimes it even turned out that the company played a discount for itself without a tangible elevator. That is, the goods would be sold the same way, at a discount, that without.
The management, seeing our work, began not only to evaluate the effectiveness of each stock, but to set the task for the departments, for example, to increase sales by 50%, and we considered the assembly of methods that would allow us to do this, since there were opportunities for combining stocks.
When does it make sense to switch to such a system?
When you know exactly why this is necessary. For example, they realized that last year 50-80% of marketing campaigns went with an unsatisfactory ROI. Or if you want to precisely and specifically justify the marketing budget not by personal feelings, but by calculations based on the analysis and extrapolation of history.
If you think with your hands and everything works, don’t touch it. But when there will be too many factors for accounting, the volumes will be of an all-Russian scale and it will be necessary to remove the influence of the human factor - go to the systems for forecasting demand and calculating the effectiveness of marketing campaigns. One of the most difficult moments is the smoothness of the transition. There are a lot of pitfalls, and all are different. Therefore, if there are questions specifically for you, ask in the comments or write to me at AlIsaev@croc.ru, I will try to help.