The vast majority of big data marketing is a waste of time

A passerby noticed a drunk man under a streetlight looking for a lost wallet. The passerby realized there was nothing on the ground, so he asked the drunk, "Where did you drop your wallet?" The drunkard pointed to the street in the distance and said, "It's right there, but I can see it better standing under the streetlight!"

We often look for answers where they are most easily reached, rather than where they are most likely to be. As marketing gradually shifts from subjective to objective, isn't a data-driven approach more like the streetlight described above?

Is it possible that even the most advanced big data methods nowadays are asking the wrong questions on the wrong data?

The wrong place to start?

When early benefit growth begins to slow, most companies tend to analyze data. One most often hears, "Let's make better use of the data we have," as if to herald the beginning of a mature company transformation. In fact, at this point in time, the period of high company effectiveness growth is officially over.

The classic question that one would initially want to address with a big data approach is: "Who are our largest customer segments?" Or "Which products are the most profitable?" The answer that can quickly be derived is that these questions are related to different regions, seasons, and many other factors. So again, we ask, "How do sales of items X,Y,Z compare in region A to region B?" Next the company uses a propensity model (PTR) for likelihood to buy analysis, cross-selling analysis, customer churn analysis, or fraud analysis, etc. In order to estimate advertising investment in different channels, one prioritizes marketing mix models in competitive marketing.

The current purpose of big data marketing is to provide companies with real-time information about who is most likely to be a customer, what channels they come through, and what products they are likely to buy at what price and at what time.

Past data isn't always useful for predicting the future?

Does massive amounts of data and data analytics really make sense? Like a drunk under a streetlight, are we being led to look for problems where it seems easiest - the data we collect is all about our customers' past sales (which is easy to do), but is it useful for understanding future sales patterns and future prospects?

Prospects analyzed through classification, clustering, or PTR patterns are standardized, rigid models. Our customers are human beings, and human beings make choices based on their rational and complex thinking, not simply finding the optimal solution to a compound problem. Take buying a car, brothers with the same genes and family environment may end up making different consumer choices. If those who are most similar show different preferences, how can we use the consumption experiences of strangers to predict our consumption tendencies? How can a consumption model generated with the consumption data of thousands or hundreds of strangers make sound consumption recommendations for us?

No single consumer can be defined in a holistic way using specific clustering or classification models. It is a complex and rapidly changing world, and these analytical models we have know very little about consumers' current preferences and consumption. Individual choices in market environments can change in an instant, and changes in product sales can force consumers to choose what is available or wait for a truly appropriate product to become available.

Promotions and discounts are one way to change the appeal of a product in a way that stimulates sales of another product while potentially causing others to stagnate. Everyone's personal financial situation is also different, and each person's decision to make each purchase is instantaneous. This makes it difficult to predict any purchase behavior.

Turning to "small data"?

The "small data" we're looking at is the ever-changing product attributes and prices. This is the data that your customers and your competitors' customers are really using to make choices. Customers are comparing and evaluating products and their offerings as much as possible. All of this ultimately determines the "shadow pricing" of the commodity in the marketplace (which is the re-evaluation of the fund's holdings by the fund manager on each valuation date, using market interest rates and trading prices).

What you want to do is maximize "willingness to pay," which is a potential client's "consumer surplus" (the difference between the highest price a consumer is willing to pay for a given quantity of a particular good and the actual market price of those goods). Consumers will then favor you or your competitors based on their preferences, depending on what attributes or services your product offers.

Analyzing customer data to reduce false estimates doesn't help your customers solve their problems; it proliferates them. Multiple permutations and combinations can exacerbate a customer's difficulty in choosing, rather than giving them a better idea of what product to choose. If you can reduce the amount of effort your customers have to put into choosing a product, they will naturally choose you.

Customers need up-to-date, reliable, valid and trustworthy advice to help them choose products. These reflect their own personal preferences and budgets, which are the two most useful.

The above is what I have shared with you about the vast majority of big data marketing is a waste of time, for more information, you can follow the Global Green Ivy to share more dry goods