跳到正文
cccskills
免费在 GitHub 公开

lead-research-assistant

Identify high-quality sales leads by analyzing your business, defining ideal customer profile, and discovering target companies with personalized contact strategies.

中文简介正在准备中。当前显示原文说明。

实测评分 未实测

查看安装方法

包含的文件(1)

  • SKILL.md7.2 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Lead Research Assistant

Use This Skill When

  • Finding potential customers for your product or service
  • Building B2B prospect lists for sales outreach
  • Identifying companies for partnership opportunities
  • Preparing target account list for sales campaigns
  • Researching companies matching your ideal customer profile
  • Discovering expansion opportunities in new markets

When NOT to Use

  • Recruitment/hiring (different purpose)
  • Consumer audience research (not B2B focused)
  • Competitive analysis only (not lead-focused)
  • Market research reports (broader scope)

Context: Lead Research Maturity

Undeveloped: Manual Google searches, unqualified leads, no strategy.

Target: Defined ICP (Ideal Customer Profile), prioritized company list, contact approaches.

Optimized: Automated prospecting, multi-channel contact data, engagement scoring, CRM integration.

Core Principle

Better leads = higher conversion. Defining your Ideal Customer Profile clearly and researching company fit systematically leads to higher-quality sales conversations and shorter sales cycles.

Instructions

Step 1: Define Your Product & Value Prop

Document your offering:

## Product Definition

**What**: [Brief product description]
**Problem Solved**: [Primary pain point]
**Ideal Customer**: [Type of company/buyer]
**Key Differentiator**: [Why you vs. competitors]

### Example (SaaS)
**What**: Cloud-based incident response platform
**Problem**: Teams waste hours on incident coordination during outages
**Ideal Customer**: DevOps/SRE teams at growth-stage tech companies
**Differentiator**: AI-suggested responses + Slack-first workflow

Step 2: Define Ideal Customer Profile (ICP)

Create comprehensive ICP criteria:

CriterionDetailsWeighting
IndustrySaaS, FinTech, HealthcareMust-have
Company Size50-500 employees (Series A-C)Strong
Revenue$5M-$50M ARRImportant
Growth StageScaling (20%+ YoY growth)Important
Tech StackUses cloud infrastructure (AWS/Azure)Nice-to-have
Pain PointsFrequent incidents, alert fatigueMust-have

Calculate Lead Score:

  • Meets all must-haves: 100 points
  • Matches 80% of important criteria: 80 points
  • Matches few criteria: 40 points

Step 3: Research Target Companies

Search approaches:

Method 1: Direct company databases

  • G2 reviews (companies in adjacent solutions)
  • Crunchbase (growth-stage companies)
  • PitchBook (funded startups)
  • LinkedIn Sales Navigator

Method 2: Market analysis

  • Industry reports
  • Funded startups in verticals
  • Recent IPOs and public companies

Method 3: Referrals and warm intros

  • Ask customers for similar companies
  • Industry events and conferences
  • Online communities (Reddit, Slack)

Build prospecting list:

## Target Companies List (Top 20)

| Company | Industry | Size | Score | Fit Rationale | Contact Info |
|---------|----------|------|-------|---------------|--------------|
| Company A | FinTech | 150 emp | 95 | Uses AWS, 40% growth, hiring engineers | [contacts] |
| Company B | SaaS | 300 emp | 88 | Series B, growing incident team | [contacts] |
| Company C | Healthcare | 75 emp | 82 | Cloud-first, struggling with alerts | [contacts] |

**Score 95-100**: Tier 1 (Hot prospects)
**Score 80-94**: Tier 2 (Good fits)
**Score 60-79**: Tier 3 (Long shots)

Step 4: Gather Contact Intelligence

For top 20 companies, collect:

  • Decision-makers: VP Eng, Head of SRE, CTO
  • Email formats (if detectable)
  • LinkedIn profiles
  • Recent news (funding, hiring, product launches)
  • Company pain signals (Twitter/LinkedIn complaints)

Tools to use:

ToolPurpose
RocketReach, Hunter.io, ClearbitEmail finding
LinkedIn Sales NavigatorDecision-maker research
Crunchbase, PitchBookCompany intelligence
G2, CapterraProduct/culture signals

Step 5: Develop Contact Strategies

Create personalized approach for each Tier 1 prospect:

## Company A — Contact Strategy

**Target Decision-Maker**: John Smith, VP Engineering

**Research Signals**:
- Recently hired 3 SRE engineers (LinkedIn update)
- Tweeted about "alert fatigue" last month
- Using PagerDuty (main competitor space)

**Contact Angle**: "I noticed you're hiring SRE talent — incident response tooling is often the bottleneck for growing teams like yours"

**Outreach Sequence**:
1. LinkedIn connection request (personalized)
2. Wait 3-5 days, then warm email
3. If no response in 1 week, try different contact
4. Follow up every 2 weeks (max 3 attempts)

**Email Template**:

Subject: Quick idea for [Company]'s incident response

Hi John,

I noticed you recently expanded your SRE team. Great to see!

We work with teams like [Competitor Company] on incident response automation — cut incident resolution time in half typically.

Worth a 15-minute call to explore if it's relevant?

[Link to calendar]

[Your name]


**Alternative Approach**:
- If LinkedIn connection fails, try Twitter/email
- If email bounces, research for direct team contact
- If cold outreach fails, look for warm intro via mutual connection

Step 6: Prioritize and Sequence

Create outreach plan:

## Outreach Campaign — Q2 2026

### Phase 1: Tier 1 (Week 1-2)
- [ ] Send to Top 5 prospects
- [ ] Track responses
- [ ] Book any interested calls

### Phase 2: Tier 2 (Week 3-4)
- [ ] Send to next 10 prospects
- [ ] Iterate messaging based on Phase 1 responses

### Phase 3: Tier 3 (Week 5+)
- [ ] Broader outreach or focus based on learnings

### Tracking
- Outreach rate: [Goal] emails/week
- Response rate target: [Goal] %
- Meeting booking rate: [Goal] %
- Win rate goal: [Goal] %

**Weekly check-in**:
- [ ] How many prospects reached?
- [ ] Response rate? (Iterate messaging if <10%)
- [ ] Meetings booked? (Target: 1-3/week)

Step 7: Refine Based on Feedback

After initial outreach:

## Learnings & Refinement

**What worked**:
- Response rate highest from [Approach]
- Best meetings from [Contact type]
- Strongest fit from [Industry/Size segment]

**What didn't**:
- Cold emails to [Title] got 0% response
- [Competitor name] messaging fell flat

**Refined ICP**:
- Upgrade from "50-500" to "100-300" employees (better fit)
- Add "Recent funding" as strong signal
- Remove "Nice-to-have" tech stack filters

**Next iteration**:
- Focus on [Top-performing segment]
- Adjust messaging to emphasize [Key benefit]
- Test new channel: [LinkedIn vs. Email vs. Community]

Output

Deliver:

  1. ICP Definition (criteria, scoring model)
  2. Target Company List (20-50 companies ranked by fit)
  3. Contact Intelligence (decision-makers, emails, signals)
  4. Personalized Outreach Strategy (approach for each Tier)
  5. Campaign Plan (sequencing, tracking, goals)
  6. Refinement Notes (based on results)

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Automate 21risk tasks via Rube MCP (Composio). Always search tools first for current schemas.

中文简介正在准备中。当前显示原文说明。

manu14357/zskills162026年7月24日 更新

Automate 2chat tasks via Rube MCP (Composio). Always search tools first for current schemas.

中文简介正在准备中。当前显示原文说明。

manu14357/zskills162026年7月24日 更新

Automate Ably tasks via Rube MCP (Composio). Always search tools first for current schemas.

中文简介正在准备中。当前显示原文说明。

manu14357/zskills162026年7月24日 更新

Automate Abstract tasks via Rube MCP (Composio). Always search tools first for current schemas.

中文简介正在准备中。当前显示原文说明。

manu14357/zskills162026年7月24日 更新

Automate Abuselpdb tasks via Rube MCP (Composio). Always search tools first for current schemas.

中文简介正在准备中。当前显示原文说明。

manu14357/zskills162026年7月24日 更新

Automate Abyssale tasks via Rube MCP (Composio). Always search tools first for current schemas.

中文简介正在准备中。当前显示原文说明。

manu14357/zskills162026年7月24日 更新

查看 manu14357 的全部技能

报告此技能的问题