Data Analyst, how to make a pricing model?

How to make a pricing model in data analysis? Are you good at it once you learn it, but you fail once you practice it? If so, you might want to read this article. The author explains in depth from four aspects: theoretical model, practical dilemma, source of the problem, and practical approach, and guides you to make a good pricing model. Recommended for those who are interested in data analysis and pricing models.

“Why is what is said in the book different from what is actually done???”

In the data field, there are many knowledge points that are "understood in books but useless in practice", so students have such doubts. Taking advantage of the beginning of the year, the system will help students answer them, which will be useful in subsequent work and interviews. Among the many theories, the pricing model is the most common one.

1. Theoretical Model

The pricing model has a standard solution in economics books.

1. Basic assumption: price is inversely proportional to sales volume
2. Calculate fixed costs (fixed inputs that are not related to sales volume)
3. Calculate variable costs (inputs that increase as sales increase)
4. Find a function of demand as price changes
5. List the formula: Profit = Revenue - Cost = Price * Sales Volume - Fixed Cost - Variable Cost

Then take the first-order derivative of the price. When the first-order derivative is equal to 0, it is the price when the profit is maximum (as shown in the figure below).

There are many similar derivatives. Because it is assumed that price is inversely proportional to sales volume, a price & sales volume change curve can be made. Correspondingly, the steeper the slope of the price & sales volume change curve, the greater the price elasticity, the product is not a rigid demand, and consumers are very sensitive to price. The smoother the slope, the more rigid the product is and the less sensitive it is to price (as shown in the figure below).

It looks very simple and clear, right? But in real work, why do few people use this thing?

2. Real Dilemma

The practical difficulty is: how to know this demand function "Q=1000-4P". The instinctive idea is: I first set a price, for example, 100 yuan, and then lower it to 90, 80, 70, 60... and then raise it to 110, 120, 130, 140, so that I can measure it, right?

Well, the idea is good, but there are 5 major obstacles in reality that make this gameplay impossible.

Level 1: Competitors

If a product sells well, it is easy for competitors to target it. It is said that e-commerce players in Hebei are so powerful that they will sell imitation clothes for 9.9 yuan and free shipping before the live broadcast ends. How can they survive? The speed of competitors will not allow the price of the products that sell well to be measured slowly.

Level 2: Consumers

If a product is really 90 today and 80 tomorrow, consumers will just sit back and wait, thinking, "It's always discounted anyway, so I might as well wait and see..." So the brand is ruined by continuous discounts, and this is the scene.

Level 3: Freeloaders

You may think of another form of testing: big data killing old customers. Test it by offering discounts in the background. This will easily attract the attention of the bargain hunters. If the product is easy to sell, there will be bargain hunting, smuggling, and random prices everywhere.

Level 4: Industry and Commerce Bureau

Needless to say, the issue of big data price discrimination is at the forefront of public opinion, and it is only a matter of time before it is criticized by the media and regulators.

Of course, there is also a hidden level 5: Marketing/Sales Department.

The marketing department will say: "I'm going to release this model to be a hit! A hit! Do you understand? Why are you talking nonsense?"

The sales department will say: "Do you have a definite price? If there is no definite price, will you be responsible if the sales volume does not increase?!!"

Not to mention the hunger marketing tactics, the first batch was limited to 100 people at 19 yuan, and if you don't buy it, the price will be raised to 199 yuan, and it is still limited to 100 people, and then it will be raised to 299 yuan, and it is still limited to 100 people... A few days later, there will be a "Cherish the hot-selling product, 199 yuan limited time return..." This routine is used in many places. Even the basic assumption that price is inversely proportional to sales volume has been broken.

╮(╯▽╰)╭ Why is there such a big difference between theory and reality?

3. Source of the problem

Note: The reason why academic research is called "academic research" is because it is not equal to real-life operations. The basic approach of academic research is to abstract the basic model framework from reality, and then combine it with reality to go deeper and deeper. So there is no need to shout "Studying is useless" or "It's all bullshit" when you encounter a problem...

The difference between pricing model theory and reality comes from two points

First: Irrational consumption can be created in reality, which violates the basic assumption of "rational man" in economics.

For example, hunger marketing mostly occurs in fan groups, which easily triggers emotional resonance and ignores the real production cost. Products of hunger marketing are often technology products, cultural and artistic products, and design products. The products are unique in design and have big gimmicks, which are easy to be withheld and sold at a high price, thus causing a rush. At this time, we should refer to the theory of marketing, not economics.

Second: In reality, most companies are just small players in the market and cannot understand the overall supply and demand situation.

For many commodities, if we count the sales volume of the entire industry, it is definitely inversely proportional to the price. However, for a specific company, the consumers it faces are very limited, and it is just a drop in the ocean of the entire demand curve. At this time, consumers are more likely to compare with similar competing products, so competing products can determine prices far more than consumer demand.

Therefore, if you want to do a good job of pricing for your own company, it is more important to combine the characteristics of your own company with the approach of your business plan, rather than studying books blindly and trying to find the only correct answer.

IV. Practical approach

In reality, there are simpler and more reliable ways to do pricing well.

First: Determine the marketing plan

The items to be determined include:

1. What is the level of product positioning?
2. What are the product’s competitors?
3. What is the planned sales volume of the product?
4. What is the cadence of product marketing?

With the grades and competitors, the product price range is basically determined.

The total sales volume and marketing rhythm determine the product pricing.

Its basic strategy can be described using a matrix model (as shown below).

Second: Redefine the cost baseline

Note that the cost verification is exactly the same as what is described in the book. As for the cost part, except for the very metaphysical cost of Internet advertising, most other costs come from the production line, so it is easy to calculate clearly.

Once you know the baseline, you can combine it with your marketing strategy to initially determine your pricing plan.

If the goal is to make small profits but quick turnover, you can start a price war.

If you are going for a high-end product, you have to sell it at a higher price.

Third: Test new product pricing

Before the product goes on the market, there is no real price & sales data, so feedback can only be obtained through testing.

There are many related methods in the field of market research, such as:

The items to be determined include:

1. Combine the function points to test which function points are valuable hard attributes and which are insensitive soft attributes
2. Finished product blind testing, all finished products are mixed together, blind testing prices to see if the product quality is up to standard
3. Brand testing: After adding the brand factor, test whether the brand + product is worth a specific price.
4. PSM test, after the brand and product are determined, elasticity test is performed to see the price increase and decrease range

I won’t list them all here, and those who are interested can look up the market research content.

In addition to research, you can also collect opinions from industry experts, dealers, and KOLs during internal product consultation, pre-ordering, and advance sales. However, most of these methods are qualitative and are not listed here.

Fourth: Product price adjustment after listing

After the product is launched, the initial price has been announced, and the subsequent price can only be adjusted based on this. However, there are many strategies for adjustment.

The price increase strategy can be hunger marketing, limited edition, or bundled combination. The price reduction strategy can be big data killing old customers, discount combination, or seasonal activities. In short, through subsequent marketing means, make up for the shortfall on the basis of the initial price. Most of these are related to marketing activity planning, and I will share them separately when I have the chance.

V. Summary

In fact, whenever book knowledge collides with public psychology, social opinion, user preferences, etc., there will be more or less deviations. Not only pricing models, but also classic models such as EOQ, AHP, and probabilistic decision-making will encounter problems. I will share them with you slowly in the future, so stay tuned.

Author: Down-to-earth Teacher Chen

WeChat public account: Down-to-earth Teacher Chen (ID: gh_abf29df6ada8)