prospect-research-integration

見込み客リサーチの統合

複数のソースから見込み客のデータを自動的にリサーチし、充実させるセールスボットの構築に精通しています。アウトリーチのパーソナライズやリードの選別に役立つ情報を収集するシステムを、開発者が作成できるよう支援することが目標です。

リサーチ統合が重要な理由

データのギャップ

What you have:
- Name: John Smith
- Email: [email protected]
- Company: Acme Corp

What you need to personalize:
- Role and responsibilities
- Company initiatives
- Recent news/triggers
- Tech stack
- Pain points
- Budget authority

リサーチ統合後

Auto-enriched profile:
- Title: VP of Engineering
- Reports to: CTO
- Team size: 45 engineers
- Recent: Announced series B
- Tech: AWS, React, Python
- Hiring: 12 open roles
- Trigger: New product launch

Now you can personalize.

データソースの統合

ソースの設定

class DataSourceManager:
    def __init__(self):
        self.sources = {}
        self.priority_order = []
        self.cache = ResearchCache()

    def register_source(self, source_id, source_config):
        self.sources[source_id] = {
            "connector": source_config["connector"],
            "data_types": source_config["data_types"],
            "rate_limit": source_config.get("rate_limit"),
            "cost_per_lookup": source_config.get("cost", 0),
            "reliability": source_config.get("reliability", 0.8)
        }

    def lookup(self, identifier, data_types_needed):
        """Lookup data from best available sources"""

        results = {}
        sources_used = []

        for data_type in data_types_needed:
            # Find best source for this data type
            source = self.find_best_source(data_type)
            if source:
                # Check cache first
                cached = self.cache.get(identifier, data_type, source)
                if cached and not cached.is_stale():
                    results[data_type] = cached.data
                else:
                    # Fetch fresh
                    data = self.fetch_from_source(source, identifier, data_type)
                    if data:
                        results[data_type] = data
                        self.cache.set(identifier, data_type, source, data)
                        sources_used.append(source)

        return {
            "data": results,
            "sources_used": sources_used,
            "completeness": len(results) / len(data_types_needed)
        }


# Source definitions
source_manager = DataSourceManager()

source_manager.register_source("clearbit", {
    "connector": ClearbitConnector(),
    "data_types": ["company_info", "contact_info", "tech_stack"],
    "rate_limit": 100,
    "cost_per_lookup": 0.10
})

source_manager.register_source("linkedin_api", {
    "connector": LinkedInConnector(),
    "data_types": ["contact_info", "work_history", "connections"],
    "rate_limit": 50,
    "cost_per_lookup": 0.05
})

source_manager.register_source("news_api", {
    "connector": NewsAPIConnector(),
    "data_types": ["company_news", "press_releases"],
    "rate_limit": 1000,
    "cost_per_lookup": 0.01
})

企業リサーチ

def research_company(company_name, domain=None):
    """Gather comprehensive company intelligence"""

    research = {
        "basic_info": {},
        "financials": {},
        "tech_stack": {},
        "news": [],
        "hiring": {},
        "social": {}
    }

    # Basic company info
    basic = source_manager.lookup(
        domain or company_name,
        ["company_size", "industry", "location", "founded"]
    )
    research["basic_info"] = basic["data"]

    # Financial signals
    financials = gather_financial_signals(company_name)
    research["financials"] = {
        "funding": financials.get("recent_funding"),
        "revenue_estimate": financials.get("revenue"),
        "growth_signals": financials.get("growth_indicators")
    }

    # Technology stack
    tech = detect_tech_stack(domain)
    research["tech_stack"] = {
        "detected": tech["technologies"],
        "categories": categorize_tech(tech["technologies"]),
        "relevant_to_us": filter_relevant_tech(tech["technologies"])
    }

    # Recent news
    news = source_manager.lookup(company_name, ["company_news"])
    research["news"] = extract_relevant_news(news["data"], days=90)

    # Hiring signals
    hiring = scrape_job_postings(company_name, domain)
    research["hiring"] = {
        "total_openings": len(hiring),
        "by_department": group_by_department(hiring),
        "growth_areas": identify_growth_areas(hiring),
        "relevant_roles": filter_relevant_roles(hiring)
    }

    return research

def gather_financial_signals(company_name):
    """Gather financial intelligence"""

    signals = {}

    # Check for recent funding
    funding = lookup_crunchbase(company_name)
    if funding:
        signals["recent_funding"] = {
            "amount": funding.amount,
            "round": funding.round_type,
            "date": funding.date,
            "investors": funding.investors
        }

    # Check for IPO/acquisition news
    corporate_events = search_sec_filings(company_name)
    signals["corporate_events"] = corporate_events

    return signals

担当者リサーチ

def research_contact(email=None, name=None, company=None, linkedin_url=None):
    """Gather comprehensive contact intelligence"""

    research = {
        "professional": {},
        "social": {},
        "activity": {},
        "connections": {}
    }

    # Professional info
    if linkedin_url:
        linkedin_data = scrape_linkedin_profile(linkedin_url)
        research["professional"] = {
            "current_role": linkedin_data.get("title"),
            "tenure": calculate_tenure(linkedin_data.get("start_date")),
            "previous_roles": linkedin_data.get("experience", [])[:3],
            "education": linkedin_data.get("education"),
            "skills": linkedin_data.get("skills", [])[:10]
        }

    # Enrich from data providers
    enriched = source_manager.lookup(
        email or linkedin_url,
        ["contact_info", "work_history"]
    )
    research["professional"].update(enriched.get("data", {}))

    # Social activity
    if linkedin_url:
        activity = get_linkedin_activity(linkedin_url)
        research["activity"] = {
            "recent_posts": activity.get("posts", [])[:5],
            "engagement_topics": extract_topics(activity),
            "content_style": analyze_content_style(activity)
        }

    # Mutual connections
    if company:
        research["connections"] = {
            "mutual_connections": find_mutual_connections(linkedin_url),
            "shared_groups": find_shared_groups(linkedin_url),
            "common_background": find_common_background(linkedin_url)
        }

    return research

トリガーイベントの検出

イベントの監視

class TriggerEventMonitor:
    def __init__(self):
        self.event_types = [
            "funding_round",
            "executive_hire",
            "product_launch",
            "expansion",
            "acquisition",
            "partnership",
            "award",
            "earnings"
        ]

    def monitor_account(self, company, domain):
        """Set up monitoring for trigger events"""

        monitors = []

        # News monitoring
        monitors.append({
            "type": "news",
            "query": f'"{company}" OR site:{domain}',
            "frequency": "daily"
        })

        # Job posting monitoring
        monitors.append({
            "type": "jobs",
            "company": company,
            "frequency": "weekly"
        })

        # LinkedIn monitoring
        monitors.append({
            "type": "linkedin_company",
            "company": company,
            "track": ["posts", "employee_changes"]
        })

        return create_monitors(monitors)

    def process_event(self, event, account):
        """Process detected trigger event"""

        # Classify event
        event_type = classify_event(event)

        # Score relevance
        relevance = score_event_relevance(event, account)

        if relevance > 0.6:
            trigger = {
                "account_id": account.id,
                "event_type": event_type,
                "event_data": event,
                "relevance_score": relevance,
                "detected_at": datetime.now(),
                "recommended_action": get_recommended_action(event_type)
            }

            # Alert for high-priority triggers
            if relevance > 0.8:
                send_trigger_alert(trigger)

            return trigger

        return None


def classify_event(event):
    """Classify event into trigger category"""

    patterns = {
        "funding_round": [
            r"raised.*\$\d+",
            r"series [a-e]",
            r"funding round",
            r"investment from"
        ],
        "executive_hire": [
            r"(hired|appointed|named).*(?:CEO|CTO|VP|Director)",
            r"joins as",
            r"new (?:CEO|CTO|VP)"
        ],
        "product_launch": [
            r"launched",
            r"announces.*product",
            r"introduces",
            r"releases"
        ],
        "expansion": [
            r"expands to",
            r"opens.*office",
            r"enters.*market",
            r"international expansion"
        ]
    }

    for event_type, event_patterns in patterns.items():
        for pattern in event_patterns:
            if re.search(pattern, event.text, re.IGNORECASE):
                return event_type

    return "other"

データの統合

プロファイルの作成

class ProspectProfileBuilder:
    def __init__(self):
        self.research_sources = []

    def build_profile(self, prospect_input):
        """Build comprehensive prospect profile"""

        profile = ProspectProfile()

        # Research company
        company_research = research_company(
            prospect_input.company,
            prospect_input.domain
        )
        profile.company = company_research

        # Research contact
        contact_research = research_contact(
            email=prospect_input.email,
            name=prospect_input.name,
            company=prospect_input.company,
            linkedin_url=prospect_input.linkedin
        )
        profile.contact = contact_research

        # Find triggers
        triggers = find_recent_triggers(prospect_input.company)
        profile.triggers = triggers

        # Synthesize insights
        profile.insights = self.synthesize_insights(profile)

        # Score profile completeness
        profile.completeness_score = self.calculate_completeness(profile)

        return profile

    def synthesize_insights(self, profile):
        """Generate actionable insights from research"""

        insights = []

        # Company + contact alignment
        if profile.contact.get("professional", {}).get("tenure"):
            tenure = profile.contact["professional"]["tenure"]
            if tenure < 6:
                insights.append({
                    "type": "new_in_role",
                    "insight": f"Recently started ({tenure} months)",
                    "opportunity": "May be evaluating new tools"
                })

        # Growth signals
        if profile.company.get("hiring", {}).get("total_openings", 0) > 10:
            insights.append({
                "type": "high_growth",
                "insight": f"{profile.company['hiring']['total_openings']} open roles",
                "opportunity": "Scaling team, may need solutions"
            })

        # Funding trigger
        funding = profile.company.get("financials", {}).get("funding")
        if funding and days_since(funding["date"]) < 90:
            insights.append({
                "type": "recent_funding",
                "insight": f"Raised {funding['amount']} {funding['round']}",
                "opportunity": "Budget available for new initiatives"
            })

        # Tech stack fit
        tech = profile.company.get("tech_stack", {}).get("relevant_to_us", [])
        if tech:
            insights.append({
                "type": "tech_fit",
                "insight": f"Using {', '.join(tech[:3])}",
                "opportunity": "Good technical fit"
            })

        return insights

リサーチのスコアリング

def score_research_quality(profile):
    """Score the quality and completeness of research"""

    scores = {}

    # Contact completeness
    contact_fields = ["title", "tenure", "previous_roles", "linkedin_url"]
    contact_complete = sum(
        1 for f in contact_fields
        if profile.contact.get("professional", {}).get(f)
    )
    scores["contact"] = contact_complete / len(contact_fields)

    # Company completeness
    company_fields = ["size", "industry", "tech_stack", "funding"]
    company_complete = sum(
        1 for f in company_fields
        if profile.company.get(f) or profile.company.get("basic_info", {}).get(f)
    )
    scores["company"] = company_complete / len(company_fields)

    # Trigger freshness
    if profile.triggers:
        most_recent = max(t["detected_at"] for t in profile.triggers)
        days_old = (datetime.now() - most_recent).days
        scores["triggers"] = max(0, 1 - days_old / 90)
    else:
        scores["triggers"] = 0

    # Overall score
    scores["overall"] = (
        scores["contact"] * 0.3 +
        scores["company"] * 0.4 +
        scores["triggers"] * 0.3
    )

    return scores

キャッシュと鮮度

リサーチキャッシュ

class ResearchCache:
    def __init__(self):
        self.cache = {}
        self.freshness_rules = {
            "company_info": timedelta(days=30),
            "contact_info": timedelta(days=14),
            "tech_stack": timedelta(days=60),
            "news": timedelta(days=1),
            "hiring": timedelta(days=7)
        }

    def get(self, identifier, data_type, source):
        key = f"{identifier}:{data_type}:{source}"
        cached = self.cache.get(key)

        if cached:
            freshness_window = self.freshness_rules.get(
                data_type,
                timedelta(days=7)
            )
            is_stale = datetime.now() - cached["timestamp"] > freshness_window

            return CacheResult(
                data=cached["data"],
                is_stale=is_stale,
                age=datetime.now() - cached["timestamp"]
            )

        return None

    def set(self, identifier, data_type, source, data):
        key = f"{identifier}:{data_type}:{source}"
        self.cache[key] = {
            "data": data,
            "timestamp": datetime.now(),
            "source": source
        }

指標

リサーチの有効性

Track:
- Profile completeness rate
- Data freshness scores
- Source reliability rates
- Cost per enriched lead
- Research to conversion correlation

Optimize for:
- >80% profile completeness
- <7 day average data age
- Cost efficiency across sources