Note ︱ Precision Marketing solutions and marketing portfolio evaluation _r︱ Precision Marketing

Often thought to climb the mountains small, can, often and really come to the starting point, Daniel, slowly footsteps to my notes to share it, please~

———————————————————————————

This note comes from the CDA-DSC-L2R course. Focus on Precision marketing solutions, marketing portfolio evaluation.

————————————————————————————————————————————

Practical Problems of precision marketing

Make full use of various technical means, push the marketing information to the more accurate audience group, thus not only save the marketing cost, but also can maximize the marketing effect. The essence of precision marketing is to find your target customers and marketing them.

Precision Marketing can solve the following problems:

To achieve solutions to expand market share identify potential customers, expand market share Activity Response Alert 1, maximize revenue within the budget 2, to respond to customers with low response probability of customer churn alert 1, confirm the loss of high probability of customers, the establishment of loss prevention measures 2, To win back the probability of high customer designated preferential marketing activities up and down Cross-sell 1, optimize the value of existing customers 2, increase customer stickiness that loyalty

————————————————————————————————————————————

Second, Precision marketing solutions

1. Predictive scheme--linear regression model

Predictive modeling is based on the user's historical information to predict their future behavior; Predictive modeling is a tool to help improve marketing activities and to achieve precision marketing for users.

2, Case one: High value target user mining

Need to do two aspects: Customer response (for each customer, predict their probability of purchase within 30 days, logistics model), customer consumption financial size (to respond to the person, forecast the amount that it will consume, OLS linear regression)

Customer Response Difficulty 1: Choice of time window

The key two time windows are needed in precision Marketing: the time window required for modeling, and the successful Response time window after marketing.

Response time window refers to the advertising marketing, consumers how many days to buy a successful marketing success, the choice of the date needs to be flexible according to different products to choose from, the general need to communicate with business companies, consultations to obtain specific date period.

Data Modeling time window refers to the time of modeling based on how long the data to speculate on the future value, you can select a week, month, year of data, specific still depends on the situation.

Modeling Difficulty 2: What metrics are needed when modeling (interpreting variables)

Different interpretation variables will have different degrees of influence in the morning, there will be very strong correlation variables, how to select these indicators is critical to modeling.

The method of selecting indexes can be used in decision tree, stepwise regression, and lasso in time dimension.

And in the selection of indicators, the reduction of dimensions is also a very important problem, the common PCA principal component analysis, but its main components in high latitude is difficult to obtain a unified, effective cognition; By contrast, the R-type clustering under clustering is easier to understand.

3, the evaluation of the results

After the model is obtained, the customer can be helped to achieve the business goal by selecting the best users under certain budget conditions according to the level of the models ' scores. Two predictions can be rendered on a single list of linked tables.

The horizontal axis represents the estimated rank of the consumption amount and the rank of whether the longitudinal axis represents the probability of purchase. Can be sorted in sequence.

————————————————————————————————————————————

Third, the marketing mix optimization

Estimating the impact of various advertising marketing actions on the different levels of sales in the case of multiple market factors intervening at the same time. Under the comprehensive influence of different advertising marketing, the effect of different marketing behavior is decomposed.

1, the marketing Mix optimization facing the problem

Marketing mix to solve the problem:

Which media has the highest conversion rate.

From the expense, such conversion rate is not efficient.

How to adjust the media mix to maximize the benefits of each payout.

How the media interact with each other and promote sales.

Our choice of advertising timing is not the best.

If we do two ad campaigns at the same time, how can we tell if one is more effective than the other?
2. Marketing performance evaluation

Two key indicators: marketing spending, ROI (relevance), original originality, and Shock Force (Impact).

Return on investment (ROI) = annual profit or annual profit/Total investment x100%

Two more critical curves: a recession curve – a measure of the lag time that a consumer gets from getting into a marketing campaign to creating a buying behavior; saturation effect – the marginal benefits of the media will decrease as investment increases

Models can be established by linear regression: sales volume = Marketing variables (TV, micro-marketing, search, newspapers, etc.) + price promotion variables + external impact variables (season, trend, competition, policy)

3. Result evaluation

3.1 Major media contributions

The main media contribution is a date-sales chart. Represents the different marketing means, different date sales situation.

From media contributions you can explore the following business logic:

1, the basis of sales, the greater the fame, many sales sources do not rely on marketing means, can naturally produce, the basis of the proportion of sales can reflect the strength of the company;

2, different times, the pros and cons of different marketing methods. From the picture, the company's online marketing effect is much better.

3.2 Media effect curve--roi curve

The media effect curve explains the relationship between advertising marketing spending and net sales

The ROI curve represents the increase in sales as investment increases. Through the ROI curve can find the best marketing portfolio, optimize the allocation of funds, so that every marketing means are allocated in the best position.

3.3 Average ROI and marginal ROI

Average roi== The average income of an ad investment that has been spent;
Marginal roi== The additional part of the advertising investment proceeds;
Find the most effective platform for additional advertising investment and optimize the distribution of advertising investment accordingly

The first column in the figure represents the different marketing tools, the second column represents the marketing budget, the third column represents the average ROI, the fourth column represents the marginal Roi, and the fifth column represents the ROI level hierarchy.

Often thought to climb the mountains small, can, often and really come to the starting point, Daniel, slowly footsteps to my notes to share it, please~

———————————————————————————