CRM Basics: The Complete Guide

What is a CRM?

A CRM is software that organizes customer data, tracks interactions, and turns both into actions. Contact details, deal stages, email history, payment status, last time you talked to them — all in one place instead of scattered across Gmail, Stripe, a spreadsheet, and your head.

The acronym stands for Customer Relationship Management. Ignore the acronym. What a CRM actually does is three things:

1. Keeps one record per customer. Not three rows in three tools that drift out of sync. One canonical record with every interaction attached.
2. Makes that record queryable. "Show me everyone who signed up in the last 30 days and hasn't paid yet" should take 10 seconds, not 10 minutes.
3. Turns the answers into actions. Filter → email campaign. Filter → pipeline stage. Filter → task list.

The ten-person spreadsheet does #1 sloppily, doesn't really do #2, and can't do #3 at all. That's the line between "CRM" and "contact list."

Why businesses need a CRM

The honest answer: most don't, at first.

If you have fewer than 20 customers, a pinned Google Doc with names, emails, and notes works. Two columns, twenty rows. Nobody needs a CRM at that scale.

The pain starts at two points:

The volume break. Somewhere between 50 and 200 customers, the human brain runs out of cache. You stop remembering who you emailed last month, whether Jane upgraded or churned, which of the seven "interested" contacts actually followed up. At that point, your memory is the bottleneck and the spreadsheet's limits (no activity log, no reminders, no queries) become blockers.

The coordination break. The moment a second person joins the revenue loop — co-founder, first sales hire, a contractor — the "in my head" parts of the customer database stop working. Two people can't both remember. You need shared state.

If you're past either of those breaks and still running on spreadsheets and inbox search, you're losing deals to simple forgetfulness. That's what a CRM solves. If you're not past either, a CRM is premature — you'll spend more time configuring it than getting value.

Types of CRM (and the ones you actually encounter)

Textbooks list three types of CRM:

• Operational CRM — automates sales, marketing, and service workflows.
• Analytical CRM — turns customer data into reports.
• Collaborative CRM — shares customer info across teams.

In practice, every real CRM does all three. The distinction is useless.

The more useful taxonomy is by who the CRM is built for:

| Flavor | Example tools | Built for | |---|---|---| | Enterprise sales CRM | Salesforce, HubSpot Sales Hub, Microsoft Dynamics | Sales teams of 20+ with SDRs handing leads to AEs. Territory management, lead routing, forecast hierarchies. | | SMB sales CRM | Pipedrive, Close, Copper | Teams of 3-20 who sell in a pipeline. Lightweight configuration, lower price. | | Personal CRM | Dex, Clay, Folk | One person managing a network of individuals, not a pipeline of deals. Relationships, not revenue. | | Modern-stack CRM | Attio, Sambandh | Internet businesses on Stripe/Supabase/Clerk who want the CRM to read from their actual stack in real time. Not a plugin — a native integration. | | All-in-one | Zoho CRM, Insightly | Covers CRM + email + project mgmt + phone. Good for companies wanting one vendor, noisy for companies wanting focus. |

The confusion happens when someone searches "best CRM" and gets recommendations across all five flavors. An enterprise CRM for a two-person bootstrapped team is wrong. A personal CRM for a $5M ARR SaaS is wrong. Pick the flavor first.

For Stripe-powered internet businesses — subscriptions, small teams, bootstrapped or early-VC — the answer is almost always the modern-stack flavor. That's who we built Sambandh for.

What to look for in a CRM

Nine features that separate useful CRMs from shelfware:

1. Native integrations with your actual stack. If your revenue is in Stripe, the CRM should read Stripe natively — MRR and churn as first-class fields, not plugin properties. Same for Supabase, Clerk, Paddle, LemonSqueezy, whoever you use.
2. A searchable contact list. Filter by 5+ dimensions at once. "Paying customers on Pro plan in the US who haven't been contacted in 30 days" should be a 3-click query.
3. A visual pipeline. Kanban board, drag to move stages, clear visual cue for stuck deals. Table-only pipelines underperform for a reason.
4. Email sync that actually works. Two-way Gmail + Outlook sync. Every sent/received email auto-associated with the right contact. No manual logging.
5. Plan-level clarity. Flat pricing you can understand in 30 seconds without talking to sales. Pay careful attention to per-contact and per-Hub tax — HubSpot's free plan becomes $1000+/mo the moment you want the features you actually need.
6. API + webhooks. If your CRM doesn't have an API, you'll regret it within a year. Automation, custom integrations, reporting outside the CRM — all need API access.
7. Audit log. Who changed what, when. Critical when things break or when you're on a compliance audit.
8. Import/export without friction. Your data is yours. CRMs that make export painful are holding you hostage — pick ones that let you download CSV + JSON in 3 clicks.
9. Fits your team size. A 47-admin-panel enterprise CRM on a 3-person team = nobody uses it. A toy CRM on a 30-person team = outgrown in 6 months. Match the stage.

Missing any one of these is usually a dealbreaker.

CRM pricing models (and the ones to avoid)

Four pricing models dominate. Only two are fair at small scale.

• Per-seat. $10–$150 per user per month. Standard for sales CRMs. Fine if your team is small and stable. Expensive fast if you grow or hire part-time contractors.
• Per-contact. Pricing tier bumps when you hit 1,000 / 5,000 / 50,000 contacts. HubSpot and ActiveCampaign use this. Punishes growth. Creates weird incentives to delete contacts to stay under a tier.
• Flat. One monthly fee regardless of contacts or seats. Much easier to budget. Sambandh's Pro ($19/mo flat) and Teams ($49/seat/mo for team collaboration) follow this model.
• Bundled Hubs. You pay for "the CRM" but the features you actually need (email, reports, sequences) are in separate paid "Hubs." HubSpot is the canonical offender. Starts free, ends $800+/mo.

The rule: for small teams, per-contact and bundled-Hub pricing is almost always traps. Flat wins. Per-seat is fine if your seat count is predictable.

Free tiers are useful but read them carefully. HubSpot's free is free forever — of contacts — but functionally crippled. Sambandh's free is 50 contacts + 1 pipeline with every feature included. Different philosophies. Pick the one where free covers your actual starting point.

When you do NOT need a CRM

Honest counter-case: CRMs solve a specific class of problem, and if you don't have that problem, you don't need a CRM.

You probably don't need one if:

• Fewer than 20 active customers. A spreadsheet is genuinely enough.
• No sales process. Product-led growth SaaS where users sign up themselves and there's no outbound — you need analytics, not a CRM. Segment, Mixpanel, or raw Supabase queries are better.
• Pure B2C transactional. DTC stores don't need a CRM per se — they need an ecommerce platform (Shopify, Lemonsqueezy storefronts) with built-in customer profiles and Klaviyo for email.
• You already have ONE tool doing CRM-shaped work well. If Notion or Airtable already runs your customer workflow and nothing is breaking, don't add a tool. Tools have cost beyond the subscription: context-switching, training, maintenance.

When one of those stops being true — pipeline gets complex, team grows, Notion starts drifting out of sync with Stripe — that's the signal.

Common CRM mistakes

Five patterns that kill CRM adoption inside small teams:

1. Buying enterprise tools for small-team needs. Configuring Salesforce for a 4-person team takes 2 weeks and the team uses 5% of the features. The tool becomes overhead.
2. Treating the CRM as the system of record for revenue. If your revenue is in Stripe, Stripe is the source of truth. The CRM should read from Stripe, not compete with it. Fighting this architecture creates data drift that only a human noticing can fix.
3. No single owner. If "everyone maintains the CRM" it decays within a month. One person is accountable for data hygiene — usually the same person who owns GTM.
4. Importing every contact from every source on day one. You get a junk contact database immediately. Better: import your active customers first. Add historical contacts selectively, tagged by source, only when you need them.
5. Never deleting data. Old leads, churned customers, defunct companies — clean them out quarterly. A bloated CRM hides current activity under noise.

The meta-mistake: treating CRM setup as a sprint instead of ongoing maintenance. It's the latter.

How to pick a CRM

Four-step decision process:

1. Figure out your type. Are you a SaaS with subscription revenue? Modern-stack CRM. A services business with long sales cycles? SMB sales CRM. A solo creator managing a network? Personal CRM. Don't evaluate tools across types.

2. List your non-negotiable integrations. Write the 3-5 tools the CRM must connect to. Stripe? Supabase? Gmail? Slack? Without those integrations, no other feature matters.

3. Try 2-3 free plans or trials, same week. Import a real subset of your actual data into each. The tool that feels obvious after 30 min wins. Don't evaluate tools you haven't actually used.

4. Check the cost of being wrong. Can you export your data cleanly? Does the tool lock you in via proprietary data formats? If switching out costs you a week of engineering, that's a high-enough switching cost to delay the decision.

If you're on Stripe and want to start concrete: try Sambandh Free. 50 contacts, all integrations, no credit card. Import from Stripe in two clicks. If it fits, Pro is $19/mo flat. If it doesn't, export takes 10 seconds.