Company logo of Crocus Information LtdCrocus Information Ltd

Business Requirements Analysis consultancy to blue chip companies

Exercise 3 description

This exercise is to give you practice in drawing a partial logical data structure diagram and describing entites.

Setting the scene

When we decomposed process 1, Sell Product and Take Order, we found out about Process 1.5, Make / receive sales calls and send emailshots, and it's this last part that needs more digging.

We're interested in the data aspects around emailshots as our existing Logical Data Model says nothing about them. This means we need to update it with the new entities and relationships we find. We're also likely to find some new attributes for entities already in the model.

So, a quick call to Bo Sibbutz (we interviewed her in exercise 1) reveals that Lou Skannon and Winnie Warmlede handle all aspects of sending the email campaigns. In speaking to LS to set up a session, he agreed and said he'd bring WW along to answer any questions about it that he couldn't.

  • First you prepare for the 'round the table' session (below)
  • Then you hold it with LS and WW, taking notes as you go
  • Next you write up the notes and pass them to LS and WW for validation
  • They correct some misunderstandings and you update the notes (below)

...and then it's over to you to draw up the partial logical data structure diagram and entity descriptions from those notes (the purpose of this exercise).

Interview Preparation

Interview preparation for understanding emailshot data entities and relationships

1 hour session booked with Lou Skannon and Winnie Warmlede on Monday 17th

From the Organisation Chart:

  • Lou Skannon - Sales Operations Manager
    • Bo Sibbutz - Sales Team Leader
    • Cole Koller - Sales assistant
    • Awd Artaker - Sales assistant
    • Winnie Warmlede - Sales assistant

Purpose

  • Update Logical Data Model with missing mailshot aspects:
    • new entities and their attributes / data items
    • relationships between them and existing entities
    • additional attributes for existing entities

Relevant analysis products for reference during the interview

(In exercise 1, question 9 and its answer are about emailshots / campaigns)

Here's the existing Logical Data Structure diagram. Note that there's nothing on it to do with emailshots (click image to see the complete data model so far)

...and here's the relevant DFD showing what we know about emailshots so far. Irrelevant detail is blurred out so we stay focused during the interview (you might choose to use a highlighter in real life) click image to see larger

Questions to cover off with Lou Skannon and Winnie Warmlede

Q1: Bo Sibbutz told us about the email addresses from the Sales Order Spreadsheet being used for the mailshots. So you only send emails to existing customers, yes? Do you get email addresses from anywhere else?

Q2: Does every customer get included in every emailshot?

Q3: Is it "opt in" or "opt out"? How does that work?

Q4: Can customers ask to receive only certain types of email?

Q5: Do you send a series of emailshots as part of the same campaign from Marketing? If so, are there any issues or problems or special actions you need to take?

Q6: Do you ever send emailshots that aren't part of one of Marketing's campaigns?

Q7: What do you do about bouncebacks when you send emailshots - you must get some, yes?

Supplemental questions arising during / after the interview

:

Ok, that's the preparation done.

As part of this exercise, familiarise yourself with what the questions are actually asking. We're trying to understand the related entities, their inter-relationships and major attributes / data items.

So each question is designed to drive out the detail needed to succeed in this, using a mixture of open-ended and closed questions.

Interview output for understanding emailshot data entities and relationships, held Mon 17th

(This has been back to Lou S and Winnie W for checking / correction)

Q1: Bo Sibbutz told us about the email addresses from the Sales Order Spreadsheet being used for the mailshots. So you only send emails to existing customers, yes? Do you get email addresses from anywhere else?

LS answer: Yes, the emailshots get sent only to existing customers, though I do include Marketing's email address and my own so we can monitor / check the campaign emails are being sent and delivered ok.

Q2: Does every customer get included in every emailshot?

LS answer: No. Some of our customers are retail outlets that only carry ladies' wear and some only carry gents' fashion. Some only take our childrens' gear. Mostly, but not always, Marketing wants male fashion offers to go only to retailers that cater to men / boys and similarly for women / girls. Other times they want every type of customer to get the emailshot.

Q3: Is it "opt in" or "opt out"? How does that work?

WW answer: When we do our regular calls to customers, we try and get / confirm their email address. At that time we record in the Sales Order spreadsheet the mailing list(s) they're to be included on, based on the types of offers they'd like to be emailed about. (Marketing call them 'campaign marketing messages'). Sometimes we get more than one email address for the same customer, as we often have more than one contact for them. One of them is always the main contact with any others really just seen by us as backups.

LS: Yes, it's 'opt in' and the macro uses the recorded opt-in preferences as part of the mailmerge process.

(Note to self: This isn't shown on the above DFD for the process - it needs updating in line with what's just been found out.)

Q4: Can customers ask to receive only certain types of email?

answer: answered above

Q5: Do you send a series of emailshots as part of the same campaign from Marketing? If so, are there any issues or problems or special actions you need to take?

LS answer: Yes. Marketing tell us what they want us to do and what messages they want to get across. Sometimes they ask us to drip feed offers to customers over two or three days.

Q6: Do you ever send emailshots that aren't part of one of Marketing's campaigns?

WW answer: No. Marketing are adamant about that. We only send emailshots on their say-so and with their message.

LS: There are exceptions. For example, we recently had a problem with our phone lines, so we quickly sent out an emailshot explaining when we expected things to be back to normal (and to let them know we were happy to receive orders by email).

Q7: What do you do about bouncebacks when you send emailshots - you must get some, yes?

WW answer: Yes, we do get them. When that happens, I go into the Sales Order spreadsheet and call the customer to check what's going on. If I need to, I get a new primary contact email address from them and update the spreadsheet with it. Then I forward the original campaign email to it.

(Note to self: Data flows for this aren't shown on the above DFD - needs updating.)

Supplemental questions arising during / after the interview

:


Now go ahead and practice your logical data modelling skills from the answers given by Lou and Winnie. You should end up with maybe six new entities that the business will recognize?

For each entity you need to identify and list its major attributes / data items.

Remember, at this stage of the investigation it's all business facing. We're only interested in the high-level business entities - the one's they'll recognise. Also, this is a logical data model, so even if there's nowhere physical that an entity appears in its entirety, it still belongs in the model.

Note that we've also discovered some flows are missing from the dataflow diagram which details the emailing process (the one in the interview preparation notes, above). This is normal.

One of the advantages of a structured method is that doing work on one analysis view (in this case, Data) can uncover aspects missing or conflicting in one of the other analysis views (in this case, Process). This is of huge benefit in getting accurate analysis results that the business trusts.


See Exercise 3 answer

There is no single right answer because we're working at a high level with business entities. For example, some business analysts may leave many-to-many relationships in place, others may break them out. See the section on Logical Data Modelling for more on many-to-many relationships.

Here's our answer - a partial Logical Data Model for you to compare yours to. Click the image to see what we came up with - but no peeking until you've tried for yourself!
click to see our answer


Exercise 1 | Exercise 2 | Exercise 3

Go to Top

Related

First things first...

To do this exercise, you'll need to have read through our Acme Fashion Supplies case study.


Get the Case Study as an ebook

Now you can carry it with you on your phone, tablet or e-reader and learn when it's convenient to you!

Client-site Business Analyst Essentials

All common ebook formats are included in the download.

If you have a discount code, enter it here:

Buy right now - only £4.50 (approx $6.50) - perfect for the daily commute!

...or you can peek inside via the ebook on Amazon

Companion Briefing Study

Our briefing study web pages, covering Requirements Analysis and Specification, are a complete guide to using a proven method and approach for business analysts. It's also available as a handy reference you can carry with you on your phone, tablet or kindle!

Client-site Business Analyst Essentials

All common ebook formats are included in the download.

If you have a discount code, enter it here:

Buy right now - only £4.50 (approx $6.50) - perfect for the daily commute!

...or you can peek inside via the ebook on Amazon