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

Axes are the things that actually decide chat tools in the field. Feature checklists (“has huddles,” “has emoji”) are omitted because they do not change the buy.

1. Time to first message

Adoption is the constraint. Frontline research keeps repeating the same filter: if a worker has no corporate email, a tool that starts with Entra ID and a training session will not be opened. WhatsApp scores 5 because it is already installed. Teams scores 1 because provisioning is the product. Zenzap scores 4: “if you can text, you can use it,” but it is still an install. Slack scores 2: channels, sidebar, and norms are a new language.

Blink and Microsoft’s own frontline docs agree on the split: Teams Frontline can work at enterprise scale with F1/F3 licenses; SMB hourly teams without IT rarely clear the onboarding step.

2. Admin and IT scale

Past a few dozen people, someone has to create groups, joiners, leavers, and permissions as a job. EU SME bands put qualitative cliffs at 10, 50, and 250 people; Dunbar’s ~150 is the informal-coordination cliff. The grader therefore loads admin weight as size grows, and loads it hard at 1,000+ where procurement and identity already exist.

Teams is 5 (Entra, Intune, admin center, F-licenses). Slack is 4 (workspaces, SSO on paid plans, Enterprise Grid). Zenzap is 3 (real admin, not a Microsoft estate). WhatsApp is 1 (the admin is whoever created the group).

3. Company-owned data

This is the axis personal messengers fail in public. The UK ICO reprimanded NHS Lanarkshire after a WhatsApp group held patient names, phone numbers, dates of birth, addresses, and clinical media for two years. The messages were not on a central server; they were on phones. NHS England BYOD guidance says leavers must have work data removed from personal devices. Meta’s own rules prevent a business from taking over an employee’s personal account. Under GDPR Article 32, the controller still owes “security of processing” wherever the data sits.

Teams is 5: Purview, eDiscovery, DLP, retention, legal hold. Slack is 4: company workspace and exports, with the serious controls on paid/Grid plans. Zenzap is 4: cloud-owned history, revoke access, hidden numbers — not Microsoft-grade eDiscovery. WhatsApp is 1: the leaver keeps the thread.

4. Tasks, sites, working hours

Deskless operations do not fail because chat lacks markdown. They fail because there is no owner, no due date, no site boundary, and no off switch after the shift. Gartner’s adjacent Collaborative Work Management market exists because conversation without work objects is not operations. Zenzap is built around that (sites, in-chat tasks, working hours) and scores 5. Slack and Teams are 3: they can attach tasks and, in Teams, Shifts, but the default unit is still a channel or a meeting. WhatsApp is 1.

This is also why hospitality, retail, logistics, construction, and “many sites” answers load this axis, and why a multi-site ops rule boosts Zenzap while cutting Slack and Teams.

5. Integrations and platform

Slack publishes 2,600+ directory apps and positions itself as the place the rest of the toolchain dumps events. That is the 2013 launch pattern (GitHub, Google Docs, Dropbox, Zendesk, Crashlytics) scaled into a marketplace. Teams is 4: Graph, SharePoint, Outlook, Power Platform — deep, but inside one vendor. Zenzap is 3: some connectors, not a developer platform. WhatsApp is 1.

Engineering culture and software-industry answers load this axis. Platform-agent answers load it further. 76% of Slack customers also use Microsoft, which is why a Microsoft-shop rule still prefers Teams: the fourth chat app is a political fight, not a feature fight.

6. Agents in the workflow

By 2026 this is a real buying criterion. Salesforce ships Agentforce inside Slack. Microsoft ships Copilot collaborative agents inside Teams and reports ~15× year-over-year growth in active agents (18× in large enterprises). The grader distinguishes three levels: suggest only, act inside the rules (tasks, nudges, handoffs), and a platform developers will build on.

Slack and Teams are 4: the platforms exist, and they assume paid AI SKUs. Zenzap is 4: agents sit in the work thread with tasks and handoffs, which is what SMB ops actually asked for. WhatsApp is 1: it is not a work-agent runtime. “Nothing yet” does not punish anyone on this axis; native/platform answers load it heavily.

7. Price you can live with

Affordability is not list-price-low. It is “will this company actually pay.” Slack Pro at $7.25/user/month annual is the highest transparent seat cost in this set and becomes absurd across hourly staff. Microsoft, after the European Commission commitments, publishes with/without-Teams deltas of about $1.60 on Business Basic up to $8.55 on enterprise SKUs. For a company already paying for Microsoft 365, Teams is nearly free to turn on — and still a 3 here, not a 5, because the bundle is the lock. WhatsApp is 5 (free). Zenzap is 4 (a seat, not a platform tax).

8. Ability to leave later

Switching cost is the hidden TCO line. Microsoft’s own Slack-to-Teams tool migrates channels and explicitly does not migrate DMs, group chats, workflows, or custom apps. Slack’s 2025–2026 API tightening made bulk history export harder for non-marketplace apps. Teams traps identity, files, and calendar in the same gravity well (portability 1). Slack is 2. WhatsApp is 4: easy to abandon as a tool, even though the archive is not a company export. Zenzap is 4: the product is sold as leave-able; it does not yet have Slack’s integration gravity.

How answers move weights

Every profile starts from a small base weight on each axis (0.2–0.4) so no metric is ever literally zero. Answers then add. The largest possible loads:

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 Teams. The IT play: Entra, Purview, Graph, Copilot. Simplicity is the tax. Portability is the worst of the four because identity, files, and calendar sit in the same vendor.
• Slack. The platform play: 2,600+ apps, search, and Agentforce. Training and per-seat price are the product. Switching cost rises with every integration.
• WhatsApp. Already on the phone, free, and trivial to leave. It is not a company system of record: no directory, no revoke-on-exit, no sites or tasks, no work-agent runtime.
• Zenzap. Messaging that feels like texting, with company-owned history, site/task structure, and in-chat agents. Weaker than Slack on ecosystem and weaker than Teams on enterprise IT and eDiscovery.

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 Workplace chats, copied from the engine:

• High or strict cyber limits eliminate personal messengers. The company must own history and revoke access.
• Offboarding needs one-click access changes, not group-by-group cleanup.
• A tiny informal team does not need Teams or Slack. Training cost is the product.
• If the team already lives in WhatsApp, cyber is low, and headcount is stable, switching is optional.
• IT-led Microsoft shops already paid for Teams. A fourth chat app is a political fight, not a feature fight.
• Many sites need chat structured by location, not a desktop collaboration suite.
• Engineering and product teams buy Slack for the ecosystem, search, and platform agents.
• Software companies buy chat that plugs into the rest of the toolchain.
• Frontline and multi-location ops need structure by site, tasks in the chat, and almost no training.
• Phone-first teams that now need company-owned data usually want familiar messaging with work controls, not a full collaboration suite.
• At 1,000+ people, incumbents with IT procurement paths usually win unless the culture is still frontline.
• If nobody wants agents and the culture is informal, Slack and Teams look like overkill.
• Custom agents and APIs favor Slack’s platform. WhatsApp is not a work-agent runtime.
• Handoffs and task follow-up need AI inside the work thread, with an owner and a due date.
• Automating joiners and leavers needs a real directory. Personal messengers cannot revoke access.
• Drafting customer replies needs the chat to reach the systems where those customers live.
• Fast hiring and churn punish tools where leavers keep the history on their personal phones.
• SMB frontline teams rarely have the IT appetite Teams assumes.
• A tight budget punishes per-seat collaboration suites. Free or near-free messaging wins until ownership hurts more than price.
• Enterprise spend makes a personal messenger look unserious. Procurement expects a platform with a contract.
• Refusing vendor lock-in cuts tools that trap history, identity, and apps inside one vendor’s gravity well.
• Companies that want to standardize will buy the platform others have to plug into. Easy-to-leave chat is not the point.

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.