The CRM Implementation Playbook: Pipedrive vs Zoho vs HubSpot for a 2–20 Person Company

A CRM implementation for a company of two to twenty people in 2026 is mostly not a software project. The platform choice matters less than the four decisions in front of it: what counts as a lead, what each pipeline stage means, which three automations run on day one, and who is responsible for the data being true. Pipedrive, Zoho CRM, and HubSpot will all hold those decisions. None of them will make them for you.

The short version of the platform question: Pipedrive if the company sells and needs a clean pipeline with the least training; Zoho CRM if you want the most fields, modules, and telephony per dollar and can live with an interface that shows it; HubSpot if marketing and sales need one database and there is budget for Professional, where the useful automation starts.

I implemented and managed all three at Pravada, a Fort Myers manufacturer, between 2020 and 2025, with pipeline automation and lead routing. WBX Creative now runs its own CRM on Supabase, which taught me more about data models than any of the three did. The bio has the rest of the record.

How to think about it

Analyst estimates of CRM project failure span roughly 18 to 69 percent, per a 2018 Harvard Business Review piece by Scott Edinger, whose diagnosis is that CRMs get used for inspection rather than for improving how the team sells. Salesforce's State of Sales 2026 research, a survey of more than 4,000 sales professionals, puts reps at 60 percent of their time on non-selling tasks. A CRM that adds to that number is a CRM nobody will use.

So treat the CRM as a written agreement the software enforces. Write the agreement first. Then pick the tool that enforces it with the fewest clicks. Everything below is in the order I do it, and skipping ahead to "buy the plan" is the most common way to end up in the 69 percent.

Step 1: Discovery, on one page

Before anyone opens a pricing page, get six answers from the owner and from whoever talks to customers:

1. Where do leads come from, by channel, per month?
2. Who touches a lead first, and who touches it second?
3. What does "won" mean? A signed quote, a deposit, a first invoice paid?
4. What does the owner want to see once a month without asking anyone?
5. What systems already exist: email (Google Workspace or Microsoft 365), phone, web forms, accounting, e-commerce.
6. How many people need to edit records, and how many only need to look?

The answers fit on one page and decide the plan. The edit-versus-look count decides seats: HubSpot's view-only seats cost nothing extra as of September 2026, which changes the math when two people sell and eight need to see the pipeline.

Step 2: The data model

Every CRM has the same four objects under different names: organizations, accounts, or companies; people or contacts; deals; and activities. Map your one page onto those four before you touch a custom field.

Then decide the minimum fields. For a small company that is usually under twenty custom fields at launch: lead source, the five UTM parameters, a service interest, a deal value, an expected close date, and a next-action date. The CRM I built for WBX stores lead source and all five UTM values on both customers and deals, because a deal that arrives six months after the first contact still needs to be credited to the channel that produced it.

Custom-field limits are a real constraint on the lower plans as of September 2026. Zoho's official comparison page lists 10 custom fields per module on Standard, 155 on Professional, 300 on Enterprise, and 500 on Ultimate. For Pipedrive, the axisconsulting.io plan listing (updated 2026-08-08; the official pricing page blocked the fetch) shows 30, 100, 300, and 500 on Lite, Growth, Premium, and Ultimate. Ten sounds like plenty until the third person asks for one.

Step 3: What should a pipeline stage mean?

A stage is something that happened, not something someone feels. "Interested" is a feeling. "Proposal sent" is an event with a date. Five or six stages is enough at this size. The WBX CRM uses five: Lead, Proposal, Negotiation, Won, Lost. Each gets one sentence of exit criteria.

Two things belong in the stage design: a required next-action date on every open deal, so "no next step" is visible as a blank, and rotting. Pipedrive's plan article lists deal rotting on every plan; the equivalent in Zoho and HubSpot is a workflow that flags deals untouched for a set number of days. A pipeline with 40 deals and no rot rule is a pipeline with 12 deals and 28 excuses.

Step 4: Three automations, then stop

The first automations should be boring:

• Routing. A new lead is assigned by territory, service, or round-robin, and the assignee gets a task due in one business day.
• Stage follow-through. Moving a deal to Proposal creates a follow-up task three business days out.
• Stale-deal alarm. No activity for 14 days on an open deal, and the owner is notified.

Everything else waits a month. Plan limits matter here: the axisconsulting.io listing puts Pipedrive at no automations on Lite and 50, 150, and 250 on the higher plans; Zoho's official comparison page has workflow rules on all paid editions and assignment rules from Professional; HubSpot's pricing page lists up to 50 workflows on Starter, 300 on Professional, and 1,000 on Enterprise; check which workflow actions Starter includes before planning routing on it.

Step 5: What should you migrate?

Every vendor has a built-in migration path, and all of them are one-way:

Platform	Tool	What to know (official help pages, accessed 2026-09-22)
Pipedrive	Import2, under Import data	Free up to 1,000,000 records; 40+ source CRMs including HubSpot, Zoho, and Salesforce; global admin only; set up users first or the admin owns every record; pipelines and custom fields can be auto-created; allow about 24 hours
Zoho CRM	Built-in API migration	From Pipedrive and HubSpot; undo or rerun at most three times; custom modules not supported; pauses if more than 5,000 records are skipped in a module
HubSpot	Smart Transfer	One-way into HubSpot from Pipedrive, Zoho, Dynamics 365, and others; Super Admin only; custom field mapping needs a Data Hub subscription

The method that avoids grief: export everything to a spreadsheet, deduplicate on email and phone, delete the columns nobody has filled in for a year, map the survivors, import into a trial account, then check 20 records by hand against the old system. Import before the users exist and the admin owns 3,000 deals.

Step 6: Integrations that pay for themselves

Email, phone, and web forms are the three that matter. As of September 2026:

Need	Pipedrive	Zoho CRM	HubSpot
Email	Two-way sync and templates from Growth, not Lite	Email insights from Standard	Gmail and Outlook inbox connection on all plans
Phone	No built-in dialer; Marketplace apps such as Aircall, JustCall, and CloudTalk	Telephony from Standard: Zoho PhoneBridge, Ring.io, Amazon Connect, and 106 others	Native calling: 500, 3,000, and 12,000 minutes on Starter, Professional, and Enterprise
Web forms	LeadBooster add-on, included on Premium and above	1 to 100 forms per module by edition (5 on Standard, 10 on Professional)	Free on all plans

The website form is the one to get right. It should write straight into the CRM with source and UTM values attached, and the same click should fire a GA4 key event, the subject of the GA4 and Tag Manager setup piece. If the form goes to an inbox and someone retypes it, the lead-source report is fiction from day one.

Step 7: Dashboards and the monthly review

Four numbers, on one screen, reviewed once a month with the owner:

1. New leads by source.
2. Open pipeline value by stage.
3. Win rate, by source if the volume allows it.
4. Days from lead to won.

At Pravada the finance side of this was unit economics, margin dashboards, and channel performance.

Step 8: Training and hygiene

Training is one hour, and the rule is one sentence: if it is not in the CRM, it did not happen. The owner runs the weekly pipeline meeting from the CRM screen and never accepts a status update that is not already in the system. That habit does more for adoption than any feature.

Hygiene is a monthly task with an owner: merge duplicates, close dead deals as Lost with a reason, and check that new records have a source. Salesforce's 2026 research found 74 percent of sales teams using AI are prioritizing data hygiene, a polite way of saying the AI features every vendor now sells only work on clean data.

Which platform for which company in 2026?

	Pipedrive	Zoho CRM	HubSpot Sales Hub
Best for	Sales-only teams that want a pipeline they can learn in an afternoon	The most fields, modules, and telephony per dollar, or companies already on Zoho apps	Marketing and sales on one database, with budget for Professional
Plans (September 2026)	Lite, Growth, Premium, Ultimate (renamed in 2025; the five-plan names are legacy)	Free (3 users), Standard, Professional, Enterprise, Ultimate	Free tools, Starter, Professional, Enterprise
Entry list price, per seat per month, annual	$14 Lite, $39 Growth, as listed by axisconsulting.io on 2026-09-22 (unverified against pipedrive.com)	$14 Standard, $23 Professional, per Zoho's edition comparison page on 2026-09-22 (the pricing page itself served Indian rupees)	Starter under $20 (the pricing page shows a $7 promo, HubSpot's blog shows $9); Professional $90 annual or $100 monthly plus $1,500 onboarding, per the official pricing page
AI, per official pages	Import assistant and Sales Assistant on all plans; AI email features on Premium and Ultimate	Zia AI agents from Professional; prediction models 2 on Enterprise, 20 on Ultimate	Agent Hub included on Starter and above; custom agents consume HubSpot Credits

My working rule: a company of two to six that sells one kind of thing goes on Pipedrive. A company of six to twenty with a service desk, inventory, or a preference for one vendor looks at Zoho. A company that sends marketing email, or plans to, and can afford Professional, goes on HubSpot for the shared database that makes the marketing worth it.

Mistakes I see

• Buying the plan before the data model. The plan decides field limits and automation counts. The data model decides which limits matter.
• A custom field for every question anyone ever asked. Fields that are not required and not reported on are empty within a quarter.
• Migrating dirty data because it is faster. It is faster for one day.
• The owner living in email instead of the CRM. If the boss asks for updates in Slack, the CRM is a chore. If the boss reads the pipeline, the CRM is the job.
• Treating Free as the plan. Fine for a trial. Lead routing, email sync, and reporting live on paid tiers.

What it costs and what to expect

Software, for five seats at September 2026 list prices with the labels above: Pipedrive Growth about $195 a month at the axisconsulting.io figure; Zoho Professional about $115 per Zoho's comparison page; HubSpot Sales Hub Professional $450 a month on annual billing plus $1,500 onboarding per the official pricing page. Re-check each vendor page before you sign.

Implementation, done properly with discovery, a written data model, migration, three automations, integrations, dashboards, and training, is listed on the WBX pricing page as a project range of $3,500 to $15,000 and up, as of September 2026. What moves it: source systems, the state of the existing data, phone and form integrations, and whether I train the team or just the owner.

Expect the first month after launch to be the hard one. That is when stage definitions get argued about and the rot rule flags deals people would rather not discuss. Both are the system working.

Where to go next

The technical services page describes how WBX Creative runs a CRM project. For what a CRM looks like once an AI assistant can read and write to it, the MCP server explainer walks through the one I built, and the stack piece answers whether a studio of one should buy any of this at all.