Cleaning CRM Data Without Slowing Sales: The Friday 20-Minute Protocol

Picture a sales manager on a Thursday afternoon. The pipeline review is tomorrow, three deals have no owner, two contacts appear twice with different phone numbers, and a rep has just told her that a sequence bounced for a third of the list. The usual answer is a cleanup project: block two weeks, export everything, fix it in a spreadsheet, re-import. (This is an illustration, not a case study.)

That project rarely survives contact with a real sales calendar. This guide offers a different approach: a 20-minute protocol, run every Friday, with a fixed checklist and a rotating owner, plus the seven fields worth standardizing first, the entry rules that stop the mess from coming back, a simple hygiene score you can track, and the checks you can automate so the 20 minutes go to decisions instead of hunting.

A note on what this is. It is a framework built from public sources (listed at the end with the date we read each one) and from common data-governance practice. We did not measure cleanup results on a CRM of our own, so there are no before-and-after claims here. Instead you get a checklist and a score that you can run on your data. Every number inside a worked example is a made-up illustration, labeled as one. This is general information, not legal, tax or financial advice.

Why big-bang cleanups fail, and what the data says

A big cleanup feels productive, and that is part of the trap. You can see the before and after, you can show the number of records fixed, and for a few weeks everything looks better. Then the CRM does what CRMs do: new records arrive with new mistakes, people change jobs, and reps keep typing the way they always typed. Within a few months the project is a memory and the problems are back.

There are four reasons this keeps happening. They come from how these projects are structured, not from a study, so treat them as reasoning:

• The source of the mess is untouched. If the entry form, the import process and the habits stay the same, the cleanup only resets the clock.
• It competes with selling. A two-week project needs people, and the people who know the data best are the reps who are trying to hit a number.
• Nobody owns it afterward. A project has an end date. Data quality does not.
• There is no definition of done. Without a measure of what good looks like, nobody can say whether the CRM is clean, so the work either never ends or stops arbitrarily.

How fast does CRM data actually go stale?

You will often see the claim that B2B contact data decays by about 30% a year. It appears in vendor blogs everywhere, and it is a convenient number for anyone selling data cleaning. The problem is that nobody seems able to point to where it came from.

Lusha, a B2B data provider, published its own measurement in 2026 and is candid about this. It counted how many contacts in a specific group, US sales VPs and C-suite executives, had a recorded employment change within 12 months. In its August 15, 2026 run, 17,719 of 140,964 contacts changed roles, a rate of 12.6%. A re-run on September 12, 2026 found 17,187 of 140,284, or 12.25%. Lusha also says it found no primary source for the 30% figure, and suggests it is closer to a two-year rate: its own two-year number for the same group was 25.67%.

Read that measurement with its limits in mind:

• It covers senior US contacts, not your whole database. Rates differ by seniority, industry and country.
• It counts only detected job changes. Lusha says this understates how often records go wrong, because phone numbers, titles and email addresses can also change without a job change.
• It comes from a data vendor, which has an interest in the topic. It is still more useful than a number with no source, because it states its method and sample.

Measure your own rate in 30 minutes

Take a random sample of 50 contacts from the records you added or last verified more than a year ago. For each one, check whether the person is still at the company and whether the email and phone still work. Divide the number of failures by 50. It is a rough figure, but it is yours, and it will teach you more than any published average. Repeat it twice a year.

Most of what a weekly cleanup involves is repetitive checking: find the duplicates, find the empty fields, find the records nobody has touched. That is the kind of work a workflow does well, and it is why we build these in n8n (may be an affiliate link) at LeenLabs. If you already know which check you want automated, you can describe it on our Request a Custom Workflow page and get a quote. If you are not sure yet, the next section gives you the routine to start with.

The Friday protocol: 20 minutes, a fixed checklist, a rotating keeper

The protocol has three parts: a fixed time, a fixed order and a rotating keeper. Fixed time because habits die when they depend on someone remembering. Fixed order because a checklist removes the need to decide what to look at. Rotating keeper because it spreads the work and the understanding, and because nobody should be the permanent janitor.

The 20-minute agenda

Rules that keep it safe

• Merge, do not delete. When two records are the same person, merge them so the history survives. Deleting loses activity and makes it impossible to audit.
• Archive, do not erase. A record with no activity for 90 days is not necessarily dead. Move it to a status that removes it from active sequences but keeps it searchable.
• Fix at the source. If the same defect appears three Fridays in a row, the problem is an entry form, an import or a habit. Fix that, not the records.
• The keeper decides in the moment. A 20-minute slot does not allow committees. Doubtful cases go to the owner, with a note, and the keeper moves on.
• Reps spend five minutes at most. Each rep fixes their own short list. The rest is the keeper’s job. If the protocol takes much more of a rep’s time, it will be dropped.
• Leave closed history alone. Do not rewrite closed-won or closed-lost records to look tidy. They are the evidence of what happened.

What "good" looks like, so nobody argues

Most cleanup efforts stall on definitions. Write down, once, what each of the following means in your company, and put it where the keeper can see it:

• Complete record: every one of the seven priority fields (next section) is filled with a valid value.
• Active record: has a meaningful activity (a reply, a meeting, a call that connected) in the last 90 days. An email open does not count.
• Duplicate: two records with the same email address, or the same full name and company. Exact matches can be merged by rule. Fuzzy matches need a person.
• Invalid: an email that bounced, a phone number that fails the format rule, or a contact the person has asked you to stop contacting.

Who can be the keeper, and how the rota works

The keeper does not need to be technical or senior. They need to be able to follow the checklist, make small judgment calls and tell an owner politely that a record needs attention. A rep, an operations person, an assistant or the owner can all do it. The only requirement is that the role moves: four to six people in rotation, each holding it for one week at a time, with a written handover at the end of the session.

• Keep a rota where everyone can see it, with the next three names and a backup for holidays.
• The outgoing keeper writes three lines: what was fixed, what is waiting and what the incoming keeper should watch.
• Skip a week only on purpose. If a Friday is missed, the next keeper starts with a 30-minute session, not a pile of guilt.
• Rotate on principle. The person who knows the data best should not be the permanent keeper, because the point is for everyone to learn what good data looks like.

Merging duplicates safely: which record wins

A merge is the only step in the protocol that is hard to undo, so decide the rules once and write them down. When two records describe the same person, one of them has to survive, and the other’s useful data has to move across. These rules cover most cases:

• The oldest record usually survives, because it carries the longest history and is more likely to be linked to deals and notes.
• Keep the most recent value for contact details, but only when it is complete and valid. A new email beats an old one. A blank beats nothing, so never overwrite a filled field with an empty one.
• Keep both records’ activities and notes. The history is the reason you are merging instead of deleting.
• If the two records have different owners, ask before merging. The owners may be working the same person from two angles.
• Merge only on a strong key. An identical email address is a strong key. A similar name is not, and two different people with the same name at the same company do exist.
• Record the merge. Write the ids of both records in the log, so a mistake can be reconstructed.

Deciding what to do with a stale record

A record with no activity for 90 days is a prompt for a decision, not a verdict. Three questions settle most cases, in order:

The seven fields to standardize first

A CRM can hold hundreds of fields. Standardizing all of them is how cleanups die. Start with the seven that drive routing, sequences, reporting and deliverability, and the rest can wait. The advice below applies whether you use Zoho CRM (may be an affiliate link), HubSpot (may be an affiliate link), Salesforce (may be an affiliate link) or another system.

Phone numbers: the format that scales

Phone fields are where free text does the most damage, because the same number can be typed five ways. The international standard is known as E.164. A number starts with a plus sign, then the country code, then the rest of the number, and the total is limited to 15 digits, with no spaces or punctuation. A US example from Twilio’s documentation is +14155552671, where 1 is the country code. Storing every number this way makes duplicate detection and calling or texting tools far more reliable. Check what your own CRM and dialer accept before you convert.

Why these seven and not others

Each of the seven answers a question someone asks every week. Email and phone: can we reach this person? Company: whose deal is this? Source: which channel is worth money? Owner: who is responsible? Stage: what happens next? Industry: which message fits? If a field does not feed a decision, it does not belong in the first round. You can audit the rest later.

Entry rules that keep the mess from coming back

Weekly cleaning treats the symptoms. Entry rules treat the cause. The goal is to make the correct way to enter data also the fastest way, so nobody has to choose between speed and quality.

A warning about required fields: every mandatory field is a place where a busy person types junk. If you force ten fields to be filled, you will get ten fields of "n/a". Require only what the next step genuinely needs, and require it at the moment it is needed.

What about the data you already have?

Entry rules protect the future. They do nothing for the thousands of records already in the system. You do not need to fix all of them, and you should not try. Sort the existing records by how much they are worth, and spend your effort in that order:

This keeps the Friday session useful: the keeper works from tier A and B lists, and tier C stops creating noise without costing anyone an hour.

Get the Automation Opportunity Scorecard

A spreadsheet that ranks your repetitive tasks by what is worth automating first. List the task, rate how often it happens, how long it takes, how error-prone it is and how many tools it touches, and the sheet orders your list. Enter your email below and we will send the download link.

Please enable JavaScript in your browser to complete this form.

Name *

First

Last

Email Name

Email *

Submit

The hygiene score: one number, tracked weekly

Without a number, cleaning is a feeling. With one, the keeper can start the Friday session by looking at a trend and asking what moved it. The score below is deliberately simple, so you can compute it in a spreadsheet from a CRM export in a few minutes.

How to use it without gaming it

• Watch the denominator. If the keeper archives every old record, Stale falls and the score rises without any real improvement. Track the number of active records next to the score.
• Show the score for the team, and each rep’s own list. Do not rank people publicly. Rankings produce arguments, not clean data.
• Set improvement targets, not a perfect score. A target of two points a month is realistic. A target of 100 invites cheating.
• Look at the parts, not only the total. A flat total can hide a rising duplicate rate behind falling stale records.
• Do not tie pay to it. The moment the score affects compensation, people optimize the score instead of the data.

Connecting the score to selling

A hygiene score is a means, not an end. Reps will only care about it if cleaner data makes their week easier, so link it to outcomes they can see. These are measures to watch, not promises; track them before and after you start, and judge the change for yourself:

• Bounce rate on email sequences. If invalid addresses fall, bounces should fall too. High bounce rates can harm your sending reputation, which affects every email you send.
• Records with no owner. Unowned leads are leads nobody follows up. This number should reach zero and stay there.
• Time from a lead’s arrival to its first touch. Clear ownership and valid contact details shorten it.
• Reps’ trust in the data. Ask your reps once a quarter, in one question, whether they believe what the CRM tells them. It is a soft measure, and it is the one that decides whether they use the system.

Share small wins in the Friday log. "Eleven duplicates merged, two of them on open deals" makes the value concrete in a way that a score alone does not.

Automating detection: duplicates, inactive records and empty fields

The first three minutes of the Friday session should not be spent hunting for problems. A scheduled report can find them for you and leave the decisions to the keeper. Here is what to automate, and what to keep human.

Two free workflows that do the detection

We publish two n8n workflows built on Zoho CRM that show this pattern. Both are written for Zoho, but the logic carries over to other CRMs, and the articles explain each step.

The first is How to Automate Zoho CRM Data Quality Audits. As described in the article, it acts as a small auditor for the leads in your Zoho CRM. On a schedule it reads every lead, checks the email address and the phone number, spots exact duplicates, adds a flag to each lead with a problem, and sends you one HTML report by email through Gmail. The article shows what each check really does, with examples, and is direct about the places where the original template needs care. One of them decides whether the workflow produces anything at all, so read that part before you run it.

The second is Flag Slowing Zoho CRM Deals with n8n. It reads deals from Zoho CRM every two hours, counts how often each deal has had activity recently compared with the period before, and picks out the deals that are slowing down. For each of those it asks an OpenAI model for a short list of follow-up tactics and posts a formatted alert to a Slack channel. The article explains how "slowing" is decided and describes a catch that decides whether the workflow does anything at all. Read it before you adapt it.

Notice what neither workflow does: it does not delete or merge anything. The first flags and reports. The second alerts. A person still decides. That is deliberate, and it is the right design for CRM data, where a wrong merge or deletion is hard to undo.

Designing your own Friday report

1. One report, not five alerts. Everything that needs the keeper’s attention should arrive in one message, in the order of the checklist.
2. Group by owner. The keeper can then send each rep a short list, not a spreadsheet.
3. Cap the length. A report with 600 lines is ignored. Show the top 25 by severity and a count of the rest.
4. Always include the score and its change. It is the first line of the agenda.
5. Log what the keeper did. A short record of each Friday is what shows the protocol is working.

If you would rather have this report built around your own CRM fields, definitions and owners, that is a straightforward request on our Request a Custom Workflow page. The free workflows library is a good place to see what else exists first.

A four-week rollout

Two rules keep the rollout honest. Change one thing at a time, so you know what moved the score. Do not add a field, a rule or a report until the previous one is working. A protocol that fits in 20 minutes is one you will still be running in six months.

When it is time to hand it to someone

Most teams can run the Friday protocol with a spreadsheet and a calendar reminder. The part that tends to get stuck is the automation: the checks that find duplicates, empty fields and stale records, and the report that puts them in front of the keeper without anyone exporting data first.

That is where LeenLabs comes in. We build custom n8n workflows and adapt existing ones, including the two Zoho workflows in this guide, to your CRM, your fields and your definitions. A request works like this:

1. You send a short form describing your CRM, the fields you want checked and what counts as stale in your team.
2. We read it and send you a quote. You see the cost before anything starts.
3. You approve the quote and pay 40% to start.
4. We build the workflow and deliver it. You pay the remaining 60% on delivery.

If you have looked for help on a freelance marketplace, you know the usual route: write a job post, collect proposals, interview candidates and manage the build yourself. That works well when you know exactly what to ask for. CRM hygiene is harder to specify, because the difficult part is not the script. It is the definitions: what counts as a duplicate, what counts as stale, who decides. A structured request puts those questions up front, so the quote you receive is for a workflow that matches your rules.

Every workflow we build is an n8n workflow, so every step and every check is visible and you can read what it does.

Not ready to request a quote? Start with the free workflows, download the Scorecard from the form above to rank your repetitive tasks, and come back when you know which one hurts most.

Frequently asked questions

How often should I clean my CRM data?

A short weekly session works better than an occasional big cleanup, because data changes continuously. Twenty minutes every Friday, with a fixed checklist, is enough for most small teams. Add a quarterly review of your picklists and rules.

What percentage of CRM data goes bad each year?

There is no reliable universal figure. The widely quoted 30% a year has no primary source that we or Lusha could find. Lusha measured about 12.6% (and 12.25% on a re-run) annual job changes for senior US contacts in 2026, which understates how often records go wrong. Measure a random sample of your own records twice a year.

Should I delete or archive old CRM records?

Archive first. Keep the record searchable, remove it from active sequences and keep the history. Delete only under a written rule, for example when a person has asked to be removed or a record is a confirmed duplicate that has been merged.

Which CRM fields matter most?

Start with email, phone, company, source, owner, stage and industry. They drive routing, sequences and reporting. Add others only when a field feeds a decision someone makes every week.

How do I get reps to follow the process without nagging?

Make the right way the fast way: picklists, defaults and duplicate warnings. Limit each rep to about five minutes a week on their own short list, and give the weekly work to a rotating keeper. Do not rank people publicly or tie pay to the score.

Can I use AI to clean CRM data?

AI can help find patterns, such as likely duplicates or inconsistent company names. It should suggest, not decide. Exact duplicates can be merged by rule. Anything uncertain, and any deletion, should go to a person.

What is a good hygiene score?

There is no universal benchmark. Compute your baseline, set an improvement target of a couple of points a month and watch the four parts, not only the total. A score is only useful when you calculate it the same way every time.

Do these steps work in HubSpot or Salesforce?

Yes. The protocol, the seven fields and the hygiene score are CRM-agnostic. The free workflows in this guide are built on Zoho CRM, so their exact nodes differ, but the logic can be rebuilt for other systems.

Sources and how we checked them

Every external figure in this guide was read on the page linked below on November 1, 2026. The worked examples, tables, formulas and thresholds are illustrations or arithmetic, and are not drawn from any source.

• Lusha, B2B Data Decay Rate: We Measured 12.6% a Year: measurement of detected employment changes among US sales VP and C-suite contacts over 12 months (August 15, 2026: 17,719 of 140,964; September 12, 2026 re-run: 17,187 of 140,284; two-year rate 25.67%), and Lusha’s statement that it found no primary source for the 30% figure.
• Twilio, What is E.164?: structure of the international phone number format, the 15-digit limit and the example +14155552671.

This article is general information and not legal, tax or financial advice. Laws and rules differ by country and state and change over time. Check with a qualified professional about your own situation.