A negative mention sitting unaddressed for a week is materially worse than one addressed in four hours. Speed isn't a vanity metric — it's the difference between an issue contained and an issue compounded. But faster-everywhere isn't the answer; the right SLA is tier-matched, the right tooling routes to the right responder, and the right workflow makes 4-hour acknowledgement repeatable. This guide covers the SLA framework, the tooling decision, the triage workflow, and the measurement loop. Pairs with negative mentions and sentiment shifts.
Before fixing, measure what's broken:
For last 90 days of mentions, capture: - Mention timestamp (when posted) - Acknowledgement timestamp (when first responded, if at all) - Resolution timestamp (when issue closed) - Platform - Severity tier (high/med/low) - Responder (who handled it) Compute: - Time-to-acknowledgement (TTA) per tier, p50/p90 - Time-to-resolution (TTR) per tier, p50/p90 - Response coverage rate (% of mentions that got a response) - Distribution by responder (who's doing the work) Typical findings: - Crisis-tier mentions get OK TTA but slow TTR - Standard mentions get OK TTR but inconsistent TTA - High % of mentions go unresponded entirely - Work concentrated in 1-2 people creating bottleneck
SLAs encode trade-off between response speed and over-investment. Calibrate per business stage:
| Tier | Definition | TTA target | TTR target |
|---|---|---|---|
| Crisis | High severity × high reach; legal/safety issues; viral negative | 2-4 hours, 24/7 | 24-72 hours initial, weeks-months full |
| High | High severity × low reach OR low severity × high reach | 24-48 hours, business hours | 3-7 days |
| Standard | Substantive but contained mentions | 3-5 business days | 5-14 days |
| Background | Drive-by mentions, low-reach social, generic complaints | 5-10 business days OR no response | N/A |
| Positive | Captured for amplification per positive mentions | 3-5 business days (thank-you) | 2-4 weeks for case-study expansion |
# Brand Mention Response SLA — v1.0 ## Tier definitions [as above] ## Coverage hours Crisis: 24/7 on-call rotation High/Standard: Mon-Fri 9am-6pm local time Background: Mon-Fri (no specific hours) ## On-call rotation (Crisis tier) - Primary: [name], +44 7... [phone] - Secondary: [name], +44 7... [phone] - Escalation: [exec name], +44 7... ## Decision rights - Crisis tier responses: comms lead approves; exec briefed - High tier responses: responding team lead approves - Standard tier responses: responder approves - Legal language: legal lead approves regardless of tier - Refunds/credits exceeding £X: finance lead approves ## Escalation criteria - Mention from named publication - Mention with 1000+ engagements - Defamation, legal threat, or regulatory mention - Pattern of related mentions emerging - Anything responder uncertain about
Under 50 mentions/month: - Google Alerts (free, basic) - Talkwalker Alerts (free, better than Google's) - Native platform notifications 50-500 mentions/month: - Brand24 (£59-£199/mo) - Mention (£41-£449/mo) - Sprout Social (£199+/mo) - Brand Mention Monitor (this platform) 500+ mentions/month: - Brandwatch - Meltwater - Sprinklr - Talkwalker (enterprise) Switching cost is non-trivial. Pick once, commit 12-24 months, review based on actual usage vs anticipated.
Alert fires →
First responder reviews (often comms / community manager):
- Read the mention
- Determine tier (severity × reach matrix)
- Determine route (support / comms / exec / legal)
- Decide response approach
Tier-appropriate action:
Crisis → Slack alert exec, draft public response, get approval, post
High → Draft response per template, peer review, post within SLA
Standard → Respond per template, no review needed (responder authority)
Background → Optional response or no action
Update tracking system:
- Mark as triaged, in progress, resolved
- Capture response text and timing
- Tag for later analytics
Resolution check:
- Did the response close the issue?
- Is escalation needed?
- Is follow-up scheduled?
Templates aren't scripts — they're starting points that responders personalise. They cut response time by 60-80% without sacrificing quality.
Hi [name], Sorry to hear you've been dealing with [specific issue mentioned]. That's not the experience we want for our customers. I'm escalating this to our [support / engineering / billing] team right now. Please DM me at [contact] and we'll get this sorted within [time commitment]. Thanks for letting us know, [Real name] [Real role]
Hi [name], Thanks for mentioning us — wanted to clarify one point: [the incorrect claim] isn't quite right. The actual situation is [brief, factual correction with link to evidence]. Happy to chat more if it'd help. [Real name] [Real role]
Hi [name], Genuinely appreciate the detailed feedback. You're right that [acknowledge what's accurate]. We've been working on [specific change] which addresses [related concern]. Your feedback helps us prioritise — [specific next step we're taking]. If you'd like to test the [improvement / new approach], I'd love to send it your way. [Real name] [Real role]
We're aware of [the situation] and taking it seriously. Our team is investigating and we'll provide a full update by [specific time]. In the meantime, customers affected can [specific action — refund path, support contact, etc]. — [Company name]
Avoid corporate-speak ("regret any inconvenience", "value our customers"). Specific, human, action-oriented language outperforms generic apology language consistently.
Monthly review metrics: - TTA per tier (p50, p90) - TTR per tier (p50, p90) - Coverage rate (% mentions responded) - Escalation rate (% bumped up tier) - Customer satisfaction follow-up (where applicable) - Response → resolution rate (did our response actually resolve the issue?) - Public perception of responses (likes, follow-up engagement) Quarterly review: - SLA hit rate per tier - Patterns in escalations (training need?) - Tool effectiveness vs cost - Workload distribution across team - Continuous improvement on response templates
Tiered alerts pipe to the right channel automatically.
Run Brand Mention Monitor →