MKT 443 Topic 3 identity resolution analysis example

Customer Relationship Management Grand Canyon University Free custom sample in 24 to 48h

This MKT 443 Topic 3 sample paper matches a composite performing arts center's patrons across ticketing, donor, email and parking systems. It was prepared for GCU MKT 443, Customer Relationship Management, catalog code MKT-443, a course in the Grand Canyon University BS in Marketing. About 126,000 records describe an estimated 71,000 people in 52,000 households. Exact email matches cut the count to about 92,000, Fellegi-Sunter scoring merges more automatically, and about 9,000 uncertain pairs go to review. A false merge is priced as costlier than a missed match, and each field gets a winning source. Merges are recorded so they can be reversed.

CourseMKT 443 Customer Relationship Management
TopicTopic 3
Paper typeIdentity resolution analysis
LengthAbout 1,000 words, 2 pages
FormatAPA 7 student paper
SchoolGrand Canyon University (GCU)
ProgramBS in Marketing
UpdatedOctober 2026

Free sample paper for MKT 443 Topic 3

1

126,000 Records, About 71,000 People: Matching a Performing Arts Center's Patrons Across Four Systems and Pricing Both Kinds of Error

[Author Name]

Colangelo College of Business, Grand Canyon University

MKT-443: Customer Relationship Management

Topic 3 Identity Resolution Analysis

[Instructor Name]

[Date]

The organization, setting and figures below are a composite written as a model document. No real employer, client, colleague or patient is described.

What this page is doingThe title gives the count the analysis reaches and its two error types. The arts center, its systems and every figure are a composite; the sources are real.
2

126,000 Records, About 71,000 People: Matching a Performing Arts Center's Patrons Across Four Systems and Pricing Both Kinds of Error

The Count That Surprised the Board

A composite regional performing arts center keeps patron records in four systems: ticketing holds about 58,000, the donor database about 14,000, the email platform about 41,000 and the parking app about 13,000, a total of about 126,000 records. The board has long described the center as serving more than 100,000 patrons. After matching, those records describe an estimated 71,000 people. This analysis shows how that estimate was reached, prices the two kinds of matching error, and states which system wins when records disagree. A single customer view, which Payne and Frow (2005) placed at the heart of information management in CRM, depends on getting this matching right. All figures are illustrative.

Step One: Exact Matches

The first step resolves the easy cases. Records in different systems that share the same email address, after converting to lower case and removing spaces, are treated as the same person. This step alone reduces about 126,000 records to about 92,000 distinct people. Exact matching is reliable when it finds a match, but it misses many: a patron who buys tickets with a work email and receives newsletters at a personal one appears as two people.

Step Two: Probabilistic Matching

The second step compares records that did not match exactly, scoring each pair on several fields: how similar the names are, whether the postal addresses agree after standardization, whether the phone numbers match, and whether the same payment card token appears in ticketing and parking. Fellegi and Sunter (1969) set out the theory behind this approach: each field's agreement or disagreement contributes evidence for or against a match, weighted by how often that agreement would occur by chance, and the combined score is compared with two thresholds. Pairs above the upper threshold are merged automatically. Pairs below the lower threshold are kept apart. Pairs between the two go to a review queue.

At the thresholds chosen, automatic merging reduces the count to about 76,000 people, and about 9,000 pairs fall into the review queue. A sample review of 500 queued pairs found that a little over half were the same person, which implies that resolving the queue would bring the total to about 71,000.

Why Some Pairs Need a Person

Some pairs are genuinely uncertain. A subscriber named Pat Moreno at one address and a donor named Patricia Moreno at the same address might be one person or two: a mother and daughter, or a married couple who share a surname and an initial. No scoring rule can settle that reliably from these fields alone. A reviewer can, often by checking giving and ticketing history together, and the review queue exists for those cases.

Pricing Both Kinds of Error

Table 1

The Two Kinds of Matching Error

ErrorWhat happensTypical cost
False merge (two people as one)Wrong receipts, giving history exposed, wrong donor statusDamaged trust and possible privacy breach; hard to undo
Missed match (one person as two)Duplicate mail, fragmented historyWasted postage; missed relationship with a loyal patron

Matching can fail in two ways, and they do not cost the same. A false merge combines two different people into one record. At an arts center, that can send one person's donation receipt to another, show one patron's giving history to a family member, or tell a development officer that a subscriber is a major donor when she is not. These errors damage trust and can breach donors' privacy. A missed match keeps one person as two records. That costs duplicate mailings and, more importantly, a fragmented view: the development office may solicit a loyal subscriber as if she were a stranger.

Because a false merge is more costly and harder to undo, the analysis sets the automatic merge threshold high and accepts a larger review queue and some remaining duplicates as the price of avoiding it.

Households Above Individuals

Couples often share a subscription but give separately and may want separate recognition. Merging them into one person would be a false merge. The analysis therefore treats households as a separate layer above individuals: each person keeps a record, and people at the same address who share a subscription are linked as a household. The 71,000 people form about 52,000 households. Ticketing and mailings can work at the household level while donor recognition stays with individuals.

Which Source Wins

When matched records disagree, a rule decides which value to keep. Postal address comes from the most recently updated record across ticketing and donations, since patrons update addresses when they buy or give. Email comes from the email platform, where patrons manage their own subscriptions. Giving history and recognition names come only from the donor database. Phone comes from ticketing. Wang and Strong (1996) noted that data users judge quality partly by whether data are believable and come from a trusted source, and naming the source for each field makes the merged record explainable when a patron asks why the center holds a particular detail.

Keeping the View Matched

Matching once is not enough, because the four systems keep creating new records. The analysis recommends running exact and probabilistic matching every night on new and changed records, sending uncertain pairs to the same review queue, and having the box office staff check for an existing patron by phone number before creating a new ticketing account. It also recommends recording every merge with the scores that justified it, so that a false merge discovered later can be reversed cleanly. A matching process that cannot undo its own mistakes would make the cost of a false merge higher still.

What the Board Should Take From This

The center serves about 71,000 people in about 52,000 households, not more than 100,000. That smaller number is better news than it sounds: it means many patrons engage in several ways, buying tickets, giving and parking, and the center can now see those relationships whole.

References

Fellegi, I. P., & Sunter, A. B. (1969). A theory for record linkage. Journal of the American Statistical Association, 64(328), 1183-1210. https://doi.org/10.1080/01621459.1969.10501049

Payne, A., & Frow, P. (2005). A strategic framework for customer relationship management. Journal of Marketing, 69(4), 167-176. https://doi.org/10.1509/jmkg.2005.69.4.167

Wang, R. Y., & Strong, D. M. (1996). Beyond accuracy: What data quality means to data consumers. Journal of Management Information Systems, 12(4), 5-33. https://doi.org/10.1080/07421222.1996.11518099

What the MKT 443 Topic 3 instructions ask for

Topic 3 of MKT-443 asks students to explain how a single customer view is built from several systems. Strong analyses show the matching steps, from exact matching to probabilistic scoring with thresholds and a review queue, grounded in record linkage theory. Students should explain why some pairs need human review, compare the costs of false merges and missed matches and set thresholds accordingly, treat households as a layer above individuals, state which source wins for each field when records disagree, and describe how matching will be kept current. Matched records should be reversible, so a mistaken merge can be undone cleanly.

Inside the MKT 443 Topic 3 example

A count that surprises the center's board opens the finished analysis. Its four systems hold about 126,000 patron records, and after matching, those records describe an estimated 71,000 people, every figure labeled illustrative. The analysis shows how that estimate was reached. Exact matches on email address resolve the easy cases. Probabilistic matching then scores pairs on name similarity, postal address, phone number and card token, and a pair above a set score is merged automatically. A band of uncertain pairs goes to a review queue, since a subscriber named Pat Moreno at one address might be two people. Households are handled as a separate layer above individuals, because a couple shares a subscription but gives separately. The analysis ends by stating which source wins for each field when records disagree.

Where the marks sit in the MKT 443 Topic 3 rubric

Faculty reward an analysis that prices matching errors. An Excellent analysis reduces 126,000 records step by step to about 71,000 people. Credit goes to explaining Fellegi and Sunter's two thresholds and the review queue for cases like Pat and Patricia Moreno. The rubric values a table comparing false merges, with exposed giving history, against missed matches. Strong analyses keep households separate from individuals and name a winning source for every field. Top marks go to nightly matching with reversible merges and a candid message to the board.

MKT 443 Topic 3 help: mistakes that cost marks

The first mark lost usually belongs to a paper treating the single view as a storage project, one big database to pour everything into, since loading four sources into one place resolves nobody. Matching rules described without thresholds leave the reader unable to judge how aggressive the merge was. A frequent gap is counting only missed matches as errors; a false merge can be worse, and for a donor it can be embarrassing in writing. Drafts that fold couples into one record because they share an address erase the separate giving the development office depends on. A precise count of unique patrons overstates what probabilistic matching can deliver, and the stronger papers report a range. Where no role owns the review queue, uncertain pairs pile up and the view quietly degrades.

Get an MKT-443 Topic 3 example written to your instructions

Send the MKT-443 Topic 3 instructions and the rubric attached in your classroom, with the systems or case your section described. We write a custom example to them, with records profiled across sources, exact and scored matching rules stated, both error types priced and survivorship set field by field, in 24 to 48 hours. The first one is free.

More MKT 443 papers and BS in Marketing samples

MKT 443 Topic 3 questions, answered

What is the difference between deterministic and probabilistic matching?

Deterministic matching joins records only when chosen fields agree exactly, such as an identical email address. Probabilistic matching compares several fields, scores how strongly they agree, allowing for spelling variants and partial addresses, and treats pairs above a threshold as the same person. The first is precise and misses a great deal; the second finds more matches and occasionally merges people who are different, which is why the example uses both.

Why is a false merge sometimes worse than a missed match?

Because a missed match leaves two incomplete records, while a false merge creates one record that is confidently wrong. At an arts center, merging two donors can put one person's gift in a thank-you letter to the other, which damages trust in a way no duplicate mailing does. Organizations handling giving or health records usually set stricter thresholds for exactly this reason.

What is a survivorship rule?

A rule deciding which value is kept when matched records disagree. One source may hold the most current address, another the legal name, another the communication preference. Survivorship can favor the most recent entry, the most trusted system or the most complete value, field by field. Without stated rules, a merge keeps whichever record the software processed last, which is rarely the best one.

Where can I find a free MKT 443 Topic 3 sample paper?

Every step of the identity analysis is printed below: the board's count, exact matches, probabilistic matching, why some pairs need a person, pricing both errors, households, which source wins, keeping the view matched and what the board should take from it. Bring your systems' record counts and we will work through them free.

Why is a false merge worse than a missed match in MKT 443 Topic 3?

A false merge can send one donor's receipt to another or expose giving history, damaging trust and privacy, and is hard to undo. A missed match mainly costs duplicate mail and a fragmented view.