CRM artesanal: mantendo ficha de cliente sem custos

Artisan CRM: client record without expensive tools

There is a black-cover notebook in the drawer of my desk at the atelier in Goiânia that started in October 2008, two months after Herreira existed. It is on its fourth binding. The first three turned into archived boxes in the corner of the room. Each page is a client: name, phone, birthday, husband or partner, children, nickel allergy or not, the piece she bought the first time, the piece she returned, what she said about coming back. There are clients in there who bought from me in 2009 and still buy. It is not a tool. It is a relationship written by hand. And it has yielded more than many expensive CRMs I have seen rise and die in friends' stores over these eighteen years. This lesson is about how to make it a system without it ceasing to be a notebook.

Counterintuitive thesis

Expensive CRM does not retain clients. A simple record filled in with discipline does. In a 2023 Bain & Company study on retention in accessible luxury retail, brands that maintain structured client records — regardless of tool — increase repurchase by up to twenty percent. The data tells part of the story. What truly retains is the specific memory the reseller demonstrates in the next conversation. And specific memory does not come from software — it comes from you having written, in your own handwriting, what mattered to that client.

Learning objectives

By the end of this lesson, you will be able to:

• Build a minimum client record with seven fields that fits on an A4 sheet or one row of a Google Sheet.
• Apply the artisan RFM framework (recency, frequency, value) without any software, classifying your current base into four actionable segments.
• Define a touch cadence per segment that respects the client and maintains presence without being invasive.
• Evaluate whether a client left your base by silence, by dissatisfaction or by life-stage shift — three reasons that demand three different responses.

Why a simple record beats expensive software

#### What Bain measured about retention

The 2023 Bain Luxury Goods Worldwide Market Study, in the section dedicated to the Brazilian accessible luxury jewelry market, compared brands that maintained structured client records with brands that relied solely on the salesperson's personal memory. The difference in twelve-month repurchase rate was twenty percent — one fifth more. The tool used did not correlate with the result. What correlated was discipline of filling.

Sebrae (Brazilian Service of Support for Micro and Small Enterprises) published a parallel study in 2023 on Brazilian micro and small retail businesses arriving at the same conclusion by another path: a simple spreadsheet, updated weekly, beats an expensive CRM abandoned in three months. The reason is behavioral: a complicated tool requires training, maintenance and team discipline — and most resellers operate alone or with one assistant. The cost of learning kills the use.

McKinsey reinforced in 2022, in its report on accessible luxury consumer behavior: simple segmentation (three to five groups) increases response to personalized communications by up to thirty percent. Fifty carefully designed segments in sophisticated CRM convert less than three manual segments well applied. Excess granularity becomes noise.

#### What fits in the seven-field record

The minimum Herreira record has seven fields. Each one was tested against the temptation to add an eighth, a ninth, a tenth — and was kept for eighteen years.

First: full name and what she likes to be called. "Maria Aparecida" but goes by "Cida." Remembering this is worth ten minutes of selling in the next conversation.

Second: phone with WhatsApp and best time. "Does not answer before ten," "avoids Sunday morning" — writing that down is respect she perceives at the second touch.

Third: her birthday and her child's or partner's. A birthday is a natural gift trigger that does not feel commercial. A child's birthday triggers a mother's gift to herself the following week.

Fourth: piece of the first purchase and date. Knowing she started with the eight-thousand-eight-hundred choker in May 2024 changes everything in the February 2026 conversation.

Fifth: allergy, nickel allergy, metal sensitivity. Critical for pieces that touch the skin. Writing it down once avoids returns and lost trust.

Sixth: one-line free observation. "Said she wants a pair of hoops for her sister's wedding in October" is the line that yields the second-half-of-the-year sale.

Seventh: last interaction — date, channel, content. "03/12 — WhatsApp — asked about rose gold solitaire earring." Without it, you reopen the conversation with "hi how are you" and lose the continuity.

Seven fields fit on a horizontal A4. They fit in one row of a Google Sheet. They fit on a notebook page. What does not fit is whatever you add on top.

#### The artisan RFM framework

Classic RFM in corporate CRM classifies clients by recency, frequency and value. The artisan version uses the same logic without software. You take your base — notebook or spreadsheet — and in one afternoon classify each client into one of four categories.

Recent warm. Bought in the last ninety days, usually buys at least once per semester, accumulated ticket above the average of your base. This is the client who deserves a touch about the new piece, with no aggressive offer.

Warm at risk. Used to buy regularly, but the last contact was between ninety and one hundred and eighty days ago. Warning sign. A reactivation touch with a question — not an offer — recovers most of them.

Valuable cold. Bought few times but with a high ticket. Probably a rare-gift client (engagement, graduation). An annual seasonal touch works better than a monthly cadence.

Discontinued cold. Disappeared more than twelve months ago. Before trying to reactivate, find out why — silence is different from dissatisfaction, and both are different from a life-stage shift (move, divorce, loss).

Comparative table: minimum client record in seven fields

Case study: reseller Cláudia's eight-hundred-row spreadsheet in Uberlândia

Context. Cláudia has been a Herreira reseller for seven years in Uberlândia. In January 2025 she had eight hundred and forty clients registered in her phone — all lost in disorganized WhatsApp Broadcast lists. Average monthly revenue: twenty-two thousand reais.

Challenge. Repurchase was falling quarter after quarter. New clients kept coming in, but old clients disappeared without notice. Cláudia had tried two paid CRMs in 2023 and 2024 — one with a monthly fee of one hundred and forty reais, another at ninety-eight — and abandoned both within ninety days due to usage complexity.

Approach. In February 2025 Cláudia opened a blank Google Sheet, with seven columns — exactly the seven fields of the minimum record. It took three Saturday mornings transferring contacts from her phone, cross-checking with order history on WhatsApp Business. She applied artisan RFM, classifying each row as recent warm, warm at risk, valuable cold or discontinued cold. She defined the touch cadence: recent warm once a month with a new piece, warm at risk once the following month with an open question, valuable cold once a quarter on a seasonal date, discontinued cold a single reactivation only with a direct question about the silence.

Result. In November 2025 — ten months later — Cláudia's average monthly revenue was thirty-eight thousand reais, seventy-three percent above the starting point. Seventy-two percent of that increase came from existing-base clients, not new ones. Tool cost: zero. Total time invested in implementation: eighteen hours spread over three Saturdays.

Lessons. Discipline of filling beat tool sophistication. Segmentation into four groups was sufficient — fifteen were not needed. The open question on reactivation ("I was thinking of you, is everything okay?") brought back more clients than a direct offer. And what the physical notebook did at the Goiânia atelier in eighteen years, the simple spreadsheet did in ten months on her phone.

Practical exercises

#### Exercise 1 — Build the record in Google Sheets (30 min)

Context. You have between fifty and five hundred client contacts scattered across WhatsApp, Instagram and your head.

Task. Open a blank Google Sheet. Create seven columns with exactly the seven fields of the Herreira record. Import your twenty most active clients from the past year and fill in everything you can remember. Mark in yellow any field where you could not answer.

Evaluation criteria.

• The seven columns are correctly named, with no extra field.
• At least fifteen of the twenty rows have five or more fields filled in.
• Yellow fields become a list of questions for the next conversation, not permanently yellow.
• You can sort the spreadsheet by birthday in less than thirty seconds.

#### Exercise 2 — Classify the base in artisan RFM (45 min)

Context. Your spreadsheet already has between one hundred and three hundred clients filled in with some level of information.

Task. Add an eighth column called "RFM segment." For each client, classify as recent warm, warm at risk, valuable cold or discontinued cold using the lesson's criteria. Count how many are in each group. Identify the biggest surprise: the group that had more clients than you imagined.

Evaluation criteria.

• One hundred percent of the rows are classified into one of the four segments.
• The count per group is in a visible cell at the top.
• You identify in one sentence which group will receive your priority action this week.
• The classification took less than forty-five minutes total.

#### Exercise 3 — WhatsApp script per segment (30 min)

Context. You defined the touch cadence, but you need the message content before sending.

Task. Write four short messages — one for each segment — that you could send tomorrow. Recent warm: new-piece offer with a personal hook. Warm at risk: open question, no offer. Valuable cold: seasonal gift trigger. Discontinued cold: direct question about the silence.

Evaluation criteria.

• Each message fits in three sentences (up to one hundred and fifty characters total when possible).
• The warm-at-risk message sells nothing — it only asks.
• The discontinued-cold message opens with vulnerability ("I noticed it has been a while") and not with an offer.
• None of the four messages starts with "Hi how are you?" without context.

Executive synthesis

Artisan CRM is not a precarious version of expensive CRM — it is the adult way to manage relationships when the operation is human, regional and trust-based. Seven fields, four segments, four cadences, four scripts. It fits in a free spreadsheet, it fits in a black-cover notebook. What does not fit is whatever you add on top out of insecurity. Weekly discipline beats sophisticated tooling every time. And specific memory — written in your handwriting or typed by you — is what the client perceives and what no software replaces.

Immediate application checklist:

• Define the seven fields of the minimum record and open a spreadsheet or notebook today.
• Import the twenty most active clients of the past year before the end of the week.
• Classify one hundred percent of the base as recent warm, warm at risk, valuable cold or discontinued cold.
• Define the touch cadence per segment (monthly, monthly, quarterly, annually).
• Write the four WhatsApp scripts and save them as ready drafts.
• Reserve one morning per week — preferably Saturday — to update the records of those who bought or messaged in the last seven days.
• Review the segment of each client once per quarter, because a valuable cold client may turn into a recent warm one over a single trip.

Next module

In the next lesson we move into WhatsApp scripts to reactivate cold clients, with ready templates for the three reasons for silence (forgetting, dissatisfaction, life-stage shift) and how to recognize each one before sending the wrong message to the wrong person.