How the grade is made

The questions

Each category asks its own quiz. The short label is what you tap; the sentence under the question is why that answer can change who wins. Security is a 1–5 scale: strongly disagree that the company must own the work, through strongly agree.

The decision model

Analyst markets and public leaderboards answer “who is winning the category?” A buyer needs “which tool fits this workforce without creating a second job?” Those are different questions. Gartner’s workstream-collaboration market is built around persistent channel-based chat, meetings, files, and an API/bot extension model. Slack and Microsoft Teams live in that market. Consumer WhatsApp does not. A large share of real companies still run work in WhatsApp anyway.

Emergence Capital’s deskless-workforce research is the reason this grader exists: about 80% of workers, roughly 2.7 billion people, do not sit at a desk, while historically almost all enterprise-software spend went to the desk-bound 20%. Ranking Slack first for a 40-person restaurant group because it “wins collaboration” is a category error.

The method is a textbook software-selection scorecard, the same shape procurement already uses:

1. Non-compensatory gates first. A true veto (for example: high cyber + personal messenger) eliminates the product. A 5 on price cannot buy back a failed gate. This is standard systems-engineering practice: mandatory vs enhancing criteria.
2. Then a weighted sum. Each remaining product has a fixed 0–5 rating on eight axes. Weights are not a committee’s opinion of “what matters in the market.” They are derived from this company’s size, industry, sites, culture, growth, security bar, budget, lock-in stance, and AI workflows.
3. Fit multipliers after that. Culture and stack matches (Microsoft shop, engineering platform, frontline ops) boost or cut a product without rewriting the underlying 0–5 ratings. Ratings describe the product. Multipliers describe this company.

Scores stay hidden until the answers are in so the result is not anchored on a public leaderboard. Any of the four products can finish last.

The formula

For each product, the engine computes a raw weighted score, then a ranking score, then a 0–100 fit grade.

raw      = Σ  weight(axis) × rating(product, axis)
maxRaw   = Σ  weight(axis) × 5
rankScore = raw × multiplier          // boosts and penalties both count
graded    = raw × min(1, multiplier)  // boosts cannot inflate a grade to 100
fit       = round(100 × graded / maxRaw)

Ranking and grading are deliberately split. A stack of “this culture loves Slack” multipliers can move Slack above Teams in the order. They cannot manufacture an A if the weighted ratings are a B. Penalties and eliminations do cut the grade. That stops the grader from becoming a home-team booster.

Letter bands are even 20-point steps, not a curve:

A 80–100B 60–79C 40–59D 0–39

An A means high fit for this profile, not “market leader.” A D can still be a famous product.

The eight metrics

CRM is not chat. The scorecard asks whether the pipeline gets updated, whether the book survives a leaver, and whether follow-ups live in the system instead of someone’s head.

How answers move weights

Small teams load time-to-first-pipeline. Field and informal teams load a pipeline people will actually update. Microsoft shops load a company-owned book next to Outlook. Software and platform answers load objects and APIs. Mid-market and up load forecast. Tight budgets and lock-in vetoes punish Salesforce theater.

Fixed product ratings (0–5)

Ratings do not change with the answers. If they did, the scorecard would be circular. They are relative to this four-product set, not to every tool on earth.

• Microsoft Dynamics 365. The Microsoft shop’s CRM. Outlook is the competitor. Implementation is the tax.
• HubSpot. Inbound CRM people actually use, with enough automation for growth. Stickier than Pipedrive; lighter than Salesforce theater.
• Pipedrive. A pipeline the team will update this week. Thin admin, thin automation, and not a system of record at 1,000 people.
• Salesforce. Depth, admin, and ecosystem. Time-to-first-pipeline and consulting cost are the product for everyone who is not already enterprise-shaped.

Fit rules and vetoes

Multipliers are the non-linear part. A 0.12 under a veto is an effective elimination. A 1.42 on an incumbent is the bundled advantage the price table already implied. Current rules for CRM management software, copied from the engine:

• A tiny team does not need Salesforce. Time to first pipeline is the product.
• Phone-first SMBs will use a CRM they can learn in an afternoon, or they will stay in the inbox.
• Frontline and multi-site teams need a CRM people will actually update, not a six-month implementation.
• Hospitality, retail, and field teams update a CRM they can learn in a week, or they do not update it.
• Software companies buy inbound and platform CRM, not a personal pipeline.
• IT-led Microsoft shops already live in Outlook. Dynamics is a procurement path, not a new religion.
• Product-led companies usually want HubSpot or Salesforce, not Dynamics by default.
• At 1,000+ people, Pipedrive is not the system of record. Salesforce or Dynamics usually win procurement.
• High cyber bars favor CRMs with SSO, audit, and enterprise admin. Lightweight pipelines get cut.
• Fast hiring outgrows a personal pipeline. HubSpot scales the inbound motion without a year of Salesforce consulting.
• Custom agents and APIs favor Salesforce and HubSpot platforms.
• SMB teams that are not already enterprise-shaped rarely finish a Salesforce or Dynamics rollout.
• A tight budget eliminates Salesforce theater. A pipeline the team will update beats a year of consulting.
• Enterprise spend expects a customer platform. Pipedrive looks like a personal pipeline, not a system of record.
• Refusing vendor lock-in cuts CRMs that trap objects, automations, and reporting inside one vendor’s data model.
• Companies that want to standardize will buy the CRM others have to plug into. Easy-to-leave is not the point.
• The motion is a list a few people work. Salesforce is theater. A pipeline they will update this week is the product.
• Inbound volume needs follow-up as a process. HubSpot is built for that motion; a personal pipeline is not.
• Field and store teams update a CRM they can learn in a week, or the book stays in their heads.
• Named accounts, forecast meetings, and a customer platform. A simple pipeline is not the system of record.

What this grader does not score

• Uptime, message latency, or independent pentest results.
• Every SKU and add-on as its own product.
• Regional data residency beyond the ownership/compliance proxy.
• Whether these products will still look like this in two years. Ratings are a snapshot of how they are actually used.

Sources

Primary public documents, regulator actions, vendor pricing and admin docs, and the deskless-workforce research that explains why a global “best chat” ranking is the wrong artifact. Product ratings for Zenzap use Zenzap’s own documentation as evidence of capabilities, then score those capabilities against the same axes as the incumbents.

1. Gartner Peer Insights — Workstream Collaboration — Market definition: persistent, channel-based conversational workspace, usually with meetings, files, and an extension model. Slack and Teams live here. WhatsApp does not.
2. Gartner Market Guide for Workstream Collaboration (quoted, 2021) — WSC tools are “intentionally designed to be extended via APIs, bots or other integration methods.” That is why ecosystem is a scored axis, not a footnote.
3. Emergence Capital — Deskless Workforce — About 80% of the global workforce, ~2.7 billion people, do not sit at a desk. Most collaboration software was built for the other 20%.
4. Emergence Capital — Software for the workforce Silicon Valley forgot — Roughly $300B a year of enterprise software spend has historically gone to desk-bound workers.
5. European Commission — SME definition — Headcount bands of <10 / <50 / <250 (Recommendation 2003/361/EC). The grader’s 2–10, 11–50, 51–200 cuts follow this, then split mid-market from 1,000+ IT procurement.
6. EUR-Lex — Recommendation 2003/361/EC — Staff headcount is the primary legal size criterion. Informal coordination fails as headcount crosses those cliffs.
7. ICO reprimand — NHS Lanarkshire (WhatsApp, 2023) — Patient names, numbers, dates of birth, addresses, and clinical images in an unmanaged WhatsApp group. Data lived on phones, not a central server. That is the ownership axis in one document.
8. Clarkslegal — WhatsApp in the workplace — UK GDPR retention, deletion, and audit fail when messages are not on company servers and when personal and work property are mixed on one phone.
9. NHS England — BYOD guidance — Leaver process must remove work data from personal devices. Subject access and FOI may require records that personal messengers cannot produce.
10. Microsoft — Teams unbundling after EC commitments — From 1 Nov 2025, published deltas: Teams is implicitly ~$1.60–$8.55/user/month inside Microsoft 365, depending on SKU. That is why Teams scores 3 on affordability for companies already on 365, not 5.
11. Slack pricing — Pro list price $7.25/user/month annual; Business+ higher; free plan caps history at 90 days and apps at 10. Affordability 2 is the paid collaboration-suite cost, not the free teaser.
12. Slack — What is Slack and how does it work? — Official figure: 2,600+ apps in the directory. 76% of Slack customers also use Microsoft. Ecosystem 5 is earned; portability 2 is the other side of the same fact.
13. Microsoft Learn — Slack to Teams Migration Tool — Migrates channel content. Does not migrate DMs, group chats, workflows, or custom apps. Switching cost is not “export a ZIP.”
14. Salesforce — Agentforce in Slack — Agents that search Slack, act in-channel, and sit on CRM data. This is why engineering/platform answers boost Slack, not WhatsApp.
15. Microsoft — Copilot collaborative agents in Teams — Agents grounded in Graph, with enterprise identity and compliance. Teams agentsFit is 4 because the platform exists; it is not 5 because it assumes Copilot spend and IT.
16. Microsoft Work Trend Index 2026 — Active agents in the Microsoft 365 ecosystem grew ~15× year over year (18× in large enterprises). Agents are now a procurement criterion, not a novelty.
17. Microsoft Learn — Microsoft 365 for frontline workers — F-licenses, Shifts, Walkie Talkie, Viva Connections. Teams can serve frontline — if IT will provision identities. The grader still penalizes Teams for SMB frontline with no IT appetite.
18. ChangeEngine — Frontline communication tools, 2026 — Deskless teams still run on noticeboards and ownerless WhatsApp groups. Slack is scored as a desk tool; the deciding filter is whether a worker without a corporate email can even log in.
19. Blink — Frontline alternatives, 2026 — Teams Frontline F1/F3 is cheap on paper and heavy in practice: email-based onboarding, notification fatigue, poor hourly-worker adoption.
20. Zenzap product documentation — Company-owned cloud history, one-click offboarding, hidden numbers, working hours, in-chat tasks, site/team structure. Ratings use this as product evidence, not as a third-party audit.
21. Software selection scorecards (WSM) — Standard practice: weight criteria, score 1–5, treat non-negotiables as pass/fail gates. The grader is that method, with weights derived from the company instead of a committee.