"data science" etiketli yazılar:

08 Kasım 2020 Pazar

Important Customer Data

Important data are not communication data, but those which help to know the customer.


One of the first assignments I gave to those who attended the CRM course in the MBA program at Istanbul Bilgi University is as follows:

For the sector you choose please specify the 5 most important data and why it is needed:

Apart from

    • Name and surname
    • Age (or Date of Birth)
    • Gender
    • Job
    • Education
    • Address / place of residence
    • GSM number / e-mail address
    • Income

Apart from those, you will briefly explain what the 5 most important data that can be obtained are and why they are important.

Important points:
When I say the 5 most important data that are OBTAINABLE, if the income or salary information of the gas purchaser will not be disclosed when asked and if you write “income is important”, the homework will be considered incomplete.

Unfortunately, some MBA participants focus only on promotion (hence communication) in studies for various sectors:

  • E-mail address: It is used to transmit all kinds of information and campaign news to the customer.
  • Phone number: It is used to transmit all kinds of information and campaign news to the customer.
  • TR ID: Used to unify customers in a customer-focused database.
  • The credit cards they use: They are used to organize campaigns with banks.
  • Marital status / Number of households: It is used to know who we can make customers except himself.
  • Media preferences: Used to position the customer’s preferred media space for marketing efforts

I doubt some of these are even ‘available’ for many industries. How many people say their “marital status” or “media preferences” when buying air tickets.

However, it is necessary to focus on the information required to get to know the customer better and thus increase the lifetime value of the customer, not on promotion.

Let’s give examples from a few different sectors:

The good and bad examples in the article are compiled from the homework of MBA participants. I thank those who prepared good examples.


  • Flight frequency
  • Frequent destinations
  • Departure place (does it always leave the same place?)
  • Total expenditure per year
  • Preferred class (Economic, Business, etc.)
  • Ticket time (last minute, discounted offers, early booking, etc.)
  • Is it through agency or personally?
  • Is he alone or flying with someone with the same last name?

There are many participants who write

    • Travel purpose
    • Job

However, we cannot know them either. It is also unlikely that we will ask every customer and expect a response. We may need to interprete over other information if it is very important to us.

We can ask ourselves “From what information do I deduce that there is a business trip?” and add that information to this list.


  • Expectations from Exercise: Weight loss / gain, fit in a bikini and wedding dress, fit body and healthy life, socialization, etc.
  • Time to Spend for Sports: How many days & hours per week and hours of arrival
  • The Sports Branch he / she wants to focus on: Fitness, swimming, group work, etc.
  • Health Status: Is there any obstacle to doing sports? Should he exercise for a specific reason?
  • Sports Background: How many years, what sports
  • Break Activities: How to use the training breaks; magazine, TV, music, chat…
  • Private trainer / Group work preference: Do you work with a trainer, prefer group work, or work alone?

In airline,  you cannot ask “Why are you traveling? Work or vacation? ” or you cannot get the correct answers. However, those who come to the gym often tell why they came.

GSM Operator:

  • Invoice amount: Cost-benefit analysis is needed to provide more reasonable offers and / or to estimate the level of income.
  • Air time and number of SMS: It is necessary to get an insight of why the customer’s preferences are phone calls or SMS and to offer better offers accordingly. It helps us learn how much of his/her time she/he spends on the phone in one way or another.
  • With / without Internet: Opportunity to cross-sell to customers based on Internet usage; online campaign offers; It is necessary to learn the most frequently used internet platforms.
  • Hours of use: The hours when the customer is busy, what hours of the day, with whom he talks
  • Numbers he speaks frequently: We can create special packages among our customers with certain numbers.
  • Phone brand: We need this information as the packages offered by the operator may vary depending on the nature of the devices. Device replacement speed also gives an idea of ​​the customer’s level of income or spending.
  • Voyage to abroad: It will be advantageous to know how often the customer visits abroad in order to avoid surprise invoices. It will allow us to provide information in advance.
  • Churn exists or not: Whether it comes from other operators, for how many years it has received service from the same institution.
  • Customer request / complaint status: It is important for behavior management towards the customer how many “problems” were opened to the customer, how many requests or complaints he made from which channels until today. What are the status of his complaints (open, solved, closed, etc.)
  • Places where he frequently speaks: School / courthouse / hospital / shopping mall / airport etc. This information is needed to make special offers using the locations where the customer is at that time.
  • Packages used: family pack, school pack, fan pack, tariffs, etc …
  • Value-added services: vehicle tracking, lost-stolen insurance, etc…
  • Sensitivity to campaigns: Personal data can be used more, it can increase income through agreements with business partners.
  • In corporate (B2B) agreements, how many people will enter the system at the same time, possible income can be estimated according to their demographic data.

I know there are more than 5 examples for each industry. Any 5 of them are considered TRUE – including other important information I couldn’t think of.

If you don’t have this information, it won’t work well, even if you have full contact information. If you do not know the customer, you can only send spam messages.

You suggest me shampoo [1] , hair curler, breast lift bra, pink face mask [2], best places to go with my deceased mother in Mothers’ Day, gifts I can buy for my deceased father for Fathers’ Day.

Your industry or homework topic may not be an airline or a gym or a GSM operator. However, by looking at the examples above, you can find out what the important information is in whatever industry you are in. You should be able to parse the “important information” that is valid even in hosting or electricity distribution companies that are thought to have a single product.

There is another important issue…

The third (or fourth) assignment of the semester is mostly on breakdowns used for segmentation. If you noticed, those who did this assignment correctly actually did the segmentation assignment as well.


Additional Materials to read:


04 Nisan 2016 Pazartesi

Veri Anlamlandırma

Birkaç ay önce, bir okulun MIS Bölüm başkanı, akademik veri görselleştirme uzmanları ile toplantıdaydık. Veri depolama işi yapan bir yabancı şirketin Türkiye’deki akademik iş ortağı olabilir miyiz diye konuşuyorduk.

Değerli Bölüm Başkanı “veri madenciliğini öylesine yaparız ki, şirketlerin bilmediği ilişkileri bile buluruz” gibi bir cümle söyledi. Yabancı şirketin genel müdürü “Artık veri madenciliğini veri biliminden (data science) saymıyoruz” dedi.

Artık veriyi ete kemiğe büründürmek, verinin arkasındaki hikayeyi bulmak… kısaca veriyi anlamlandırmak, veri bilimi sayılıyor.


Veriyi anlamlandırmanın bir ustalık işi olduğunu belirtmiştim.

Sizin tecrübeniz arttıkça, veriyi daha iyi anlamlandırmaya başlarsınız. Şöyle ki…

1988’de Bireysel Bankacılık’a ilk girdiğimde kayıtlarında “protestolu senet” olan kişilere sonradan ödemeyi yapmış olsalar bile kredi kartı verilmezdi. Krediler Departmanı, “Borcuna sadık değil” anlamını çıkartırdı.

1993’de bankadan ayrıldım. 1994 krizi sırasında leasing sektöründeydim.

1997 senesinde Bireysel Bankacılık’a, kredi kartlarına geri döndüm. O da ne? Protestolu seneti olan müşterilere – eğer sonradan borçlarını ödemişse – rahatça kredi kartı veriliyor. Ama şirketin kuruluş yılı 1994’den önce ise ve hiç protesto kaydı yoksa, daha fazla araştırma yapılıyor.

Anlam değişmişti. Protestolu senedini sonradan ödeyen kişi, “bir krizi atlatabileceğini ispatlamış” sayılıyordu. Hiç olumsuz kaydı bulunmayanlar ise, “acaba şirket vardı ama iş mi yapmadı“, “göstermelik tabela şirketi mi” diye daha kuşkuyla karşılanıyordu.


Her vesileyle anlatırım. Anlamlandırma, yukarıdaki resimdeki gibi “Taş ıslaksa yağmur yağıyor, taş beyazsa kar yağıyor” değildir.

Verinin arkasındaki hikaye, çoğunlukla sokaktaki adamın zaten bildiğinden düşündüğünden farklıdır. Bu nedenle tecrübe ve ustalık gerektirir.


Not: Bu yazı,  Uzaktan CRM Eğitimi sitesindeki diğer veri anlamlandırma yazılarına eklemedir.