Table of Contents

#1. Step 1: Define Database Functions
#2. Step 2: Determine Information Requirements
#3. Step 3: Decide how to collect information
#4. Step 4: Choose Technology
#5. Step 5: Choose a hardware platform
#6. Step 6: Fill the database (sources)

A CRM database is all customer information that you collect for use in your customer relationship management (CRM) system. For example, such a database may include the following customer information:

• contact name, phone number, name and email address, links to accounts on social networks, Skype descriptor;
• activity records, such as recent visits to the website;
• date of first contact, sale and last conversation;
• orders made by the client, their cost;
• sources of leads and their results;
• names of children, pet nicknames, hobbies and other personal data that will help build friendly relations with the client.

With this data, a business can conduct quality marketing campaigns, build loyal relationships with customers and ultimately generate new sales. In other words, CRM is the sales engine, and the customer database is the fuel that makes it work.

Today we will look at how to produce such "fuel", that is, how to create a CRM database that will bring profit to your business.

Step 1: Define Database Functions

There are four types of CRM: operational, analytical, strategic and collaboration. Operational CRMs use customer personal data to streamline operations (call-down, email distribution, etc.) Analytical gather information about customer activity for developing contact scenarios and scheduling contacts.

Strategic use of data about customers, competitors, markets, market offers and other information to determine the target audience and search for the best marketing strategy. Combined usually use operational and analytical data, for example, in one of the following ways:

• a customer support worker needs to access the customer’s transaction records when a telephone request is received;
• the hotel receptionist needs information about the past activity of the client in order to reserve a room of his preferred type (view from the window, for smokers or non-smokers);
• marketers need customer contact and personal data to do newsletters, run advertising campaigns and conduct surveys;
• a bank employee must check the status of the account in order to find out if the client has reached the credit limit.

Operational data is usually stored in databases like OLTP (online transaction processing), analytical data - in databases like OLAP (online analytical processing). OLTPs are designed to quickly deal with small-sized transactions, but with a large flow, so the data in the OLTP must be current and very accurate (quotes on financial markets, invoice status). Information in such databases usually comes from internal sources of the company.

OLAP can work with less relevant data, since the information in them is a generalized and restructured part of the OLTP, which is necessary to perform certain analytical tasks. The analytical database can be filled from both internal and external sources.

Step 2: Determine Information Requirements

To answer the question: "What information is needed" best of all will be able people who communicate and interact with customers to service, marketing and sales, as well as those who make strategic decisions in the CRM system.

So, a direct-marketer planning an advertising campaign through an email newsletter will want to learn CTR and CTOR for past advertising campaigns with a division by target market, performer and offer. Such a marketer also needs email addresses, a greeting preference (name, dear) and information about what is best to use: html or plain text.

A senior manager who makes strategic CRM solutions needs a different set of data: target audience, market segmentation, competitor analysis, and more. In other words, the choice of data to be collected depends on how and for what it will be used.

In most cases, this is the following information:

1. Contact details. This usually includes information about who is the main contact person (name) and who else (other names) is involved in making purchasing decisions - what are their roles; addresses of client accounts, phone numbers, emails, accounts in social networks, instant messengers and postal addresses.
2. Contact history. Information about who, when and how communicated with the client - dates and tools (phone, email, instant messenger). Description of communication results and contact person’s notes.
3. Transaction History. Description of purchases (price and date of purchase) and reactions to offers - what they offered to buy and what was bought from it.
4. Current pipeline. Analytics on current opportunities (status of the offer) for each specific client. For example, what is the probability that he will buy the proposed product or service (just like that or at the promotional price), expressed as a percentage: 10, 20, 30 ... 90% success. Some CRMs allow you to predict such probabilities for various marketing tools.
5. Opportunities. Unlike the “contact history”, which reflects the past, this section is focused on the future. It records potential opportunities that have not yet been proposed and implemented.
6. Products. It indicates what products and services the client has bought. Whether there were any problems associated with these purchases (for example, there was damage during delivery or an incomplete set), how they were resolved.
7. Communication preferences. Description of the most effective way of communication or the one that the client prefers: email, telephone, specific instant messenger, social network. Here you also need to specify the preferred time and treatment option (name, dear, Mr. or Mrs.).

When collecting customer data, it is important to remember that they must be relevant. Many companies allow customers to take part in the collection and updating of personal data or to prohibit the service from doing so. In addition, customers are also allowed to choose which advertising content they will accept. For example, Amazon customers can choose distribution options: new purchases, special offers, purchase conditions, surveys, notifications about renewals of subscriptions to magazines, information about Amazon partners.

Step 3: Decide how to collect information

Information for databases can be obtained from internal and external sources. The basis of most CRM is internal data. The volume of such databases depends on the degree of contact of the company with the client and legislation (some data can be collected only with the permission of the source). The collected information can be stored and used for own needs, as well as sold (not all) through partners, agents and distributors.

Internal data can be collected from various sources:

1. The marketing department may collect data on customer profiles, attraction channels, purchases made, product requests, size, and market segment.
2. The sales department has information on the purchase history (time, amounts, regularity), contact details, billing addresses, purchase criteria (warranty, additional services, promotions), reaction to promotional offers and customer preferences of customers.
3. The support service keeps records of service history, customer satisfaction levels, service conditions, complaints about problems and how to solve them.
4. The finance department has access to data on payment history, payment mechanisms, payables and receivables.
5. The webmaster can collect customer flow data.

External data can be taken from a wide variety of sources, from buying on Facebook to using government databases on people and businesses. You can also order a data collection service for a specific target audience from consulting and marketing research agencies. Data from external sources can be divided into three groups:

Compiled list data collected by bureau lists or list providers. Such companies collect information from various home, business and government sources. For example, in tax reporting, annual reports of enterprises and warranty cards. This information can be both bought and rented for a certain period.

This service is useful if you, for example, want to move from sales of sportswear to selling dance costumes and accessories. In this case, the bureau lists for your order will look for such data:

• students enrolled in dance or school courses;
• people who bought tickets for dance performances;
• people who in surveys indicated to interest in dancing;
• subscribers to specialized magazines, websites, mailings;
• dance courses.

Census data obtained from state censuses. Depending on the country and specific region, the list of available information and its relevance may be different. So, in the US, a census is held every ten years, and thanks to it, you can get data on:

• average household size;
• average income by region and group;
• percentage of people who have graduated;
• average home and mortgage value;
• ethnic affiliation;
• family affiliation.

Simulated data, generated by third parties based on data from their sources. This is a specially processed data that helps to make analytical and strategic decisions. An example of data processing is the PRIZM customer classification scheme developed by Claritas. PRIZM divides all postal codes in the country into 72 clusters depending on the answers of people to various questions about lifestyle and demographic questions.

A more complex example is the European Value Study Project and the World Value Study covering 100 countries and 90% of the world's population. With the help of empirical and subjective data (from questionnaires), scientists study cultural differences, global sociocultural and political trends, religious beliefs, and much more.

Step 4: Choose Technology

Hierarchical database. This is a rather old way (model) of storing and cataloging information that was popular between the 1960s and 1980s. Such a model can be represented in the form of a family tree, where a child can have only one parent and as many children as he wants. Access to information in such databases can be obtained by moving from the upper to the lower levels and only in this way.

Network databases. Information is organized by using fields and sets. Fields link categories into a hierarchical structure, as in the family tree mentioned above. Sets define a one-to-many relationship, that is, link categories outside the hierarchical structure of fields. For example, the book of the famous physicist Michio Kaku "Einstein’s Cosmos" in the store's catalog can be attributed to the sections "Popular science literature", "History" and "Biography".

Relational database. Stores information in the form of two-dimensional tables consisting of columns and rows. In such tables there is one or several fields that are the unique identifiers of each record - the primary key. This key helps to establish links between different tables. In CRM, it is usually assigned to each client or groups of the target audience.

Relational databases are a standard for CRM applications. How they work can be represented by a simple example when you buy a product in an online store and leave your personal data to the seller: name, phone number, address, delivery method, credit card number.

In this case, the store enters information in several databases (tables):

1. In the “Orders received” table, the order, amount and preferred shipping method are registered.
2. Inventory table includes information on inventory reduction.
3. The payment and card number are recorded in the “Payment” table.
4. In the “Customer” table, a customer record is created indicating the source of the lead, contacts, product purchased, purchase amount, etc.

Step 5: Choose a hardware platform

Several factors influence the choice of hardware platforms:

1. Database size. Personal computers and cloud services can handle large volumes of information and their processing. If the database is small, it can fit into a regular smartphone.
2. Ability to integrate. If a company already uses data collection or CRM systems, the created database should be integrated into the existing system, including working on the same platforms.
3. User Preference. Most office applications are compatible with Windows and Mac OS when it comes to desktop devices; Android and iOS, if it is a mobile device. However, if most CRM database users use Linux or Blackberry, choose these platforms.
4. Popularity. The platform may not have a large number of users and support from developers of important programs for office, marketing and CRM, but if it is in trend and its popularity is constantly growing, you need to use it too. So it was with the iPhone in 2008.

Step 6: Fill the database (sources)

Purchased data. Perhaps the most common way to populate and grow databases. Using it, you usually pay for information about each client, and the more segmented the data, the more expensive they are.

Marketing content. Collect information using your own relevant blogs, sites, groups in social networks and channels on video hosting sites. In this case, data is collected by subscription or registration.

Directories. Using data from reference books, online directories, listing users on competitors' sites, “yellow pages” of chambers of commerce and the like. If these are online catalogs, then this method of data collection is simple, but tedious, if paper-based is tedious and time consuming.

Third-party mailing lists. Use a third-party CRM database to send emails on your behalf. So you can save a lot if you need one-time mailings, and at the same time there will be no problems with compliance with the laws on personal data.

Social networks. Facebook, Twitter, Instagram(how to build a mobile app), LinkedIn(how to build a social networks) and other social networks are a great tool for finding new customers, since about 80% of the world's population have an account with at least one of them.

Events. Organize any events and collect participant data. Quite an old way, but it works fine in the era of the Internet, especially if it is a specific target audience (art, luxury, big business, etc.).

But how to hire the best developers?