Video Summary - The Sourcing Wizard: Dean Da Costa's Advanced IT Recruiting Strategies Screenshare

Business-focused summary (IT recruiting sourcing playbook)

Core idea: “Sourcing is where it all starts”

Recruiting success depends first on finding candidates—not just phone screening or closing.
Dean frames sourcing as a lifecycle:
1. start with research
2. build increasingly precise searches
3. contact candidates only after improving fit and contactability

Source construction framework (research → search string → targeting → contactability)

1) Research the target skill/term (build synonym set)

Use a “word lookup” step to confirm meaning and collect synonyms.
Example: looking up “Java” and expanding to variants like:
- JavaX
- J2SE
- J2EE

Framework component

Synonym expansion improves search-string coverage across platforms where titles/keywords vary.

2) Translate skills into search intent across platforms

Identify candidate repositories first (e.g., social profiles, resume/PDF sites, code platforms).
Then adapt search logic to each platform’s quirks, such as:
- how “where they work” fields appear
- differences in resume formats and document layouts

Playbooks / tactics demonstrated in the screenshare

A) “X-ray” + proximity search (advanced Boolean-style targeting)

Example: x-raying into Facebook for “Java developer”, using signals near employment/current-title areas.
Query logic: search for the phrase within a window (he demonstrates a “10 words before/after” concept around job keywords).
Warning: small changes in keyword combinations can significantly change results:
- generic titles can increase volume
- overly specific terms can reduce results (including “robot/too-specific” effects)

Operational tactic

Start broad enough to generate candidates, then tighten with requirements.

B) Resume-first searching using images

He recommends searching images because many platforms host real resumes/CVs as documents.
Tool behavior: often automatically looks for resume/CV/bio/profile terms—so you don’t need to spell them out manually.
Noise control: filter by document color characteristics (he suggests black/white filters to reduce junk).

Actionable recommendation

Use image/PDF searching to capture candidates who never list their full skill keyword in plain text titles.

C) LinkedIn constraints → compensate with email-enrichment scraping

Key limitations/claims

Free/limited LinkedIn access returns up to 1,000 results and tends to focus on closer connections.
Searching LinkedIn alone limits reach.
LinkedIn IM outreach is not ideal if you want direct contact.

Process demonstrated

Search LinkedIn for candidate titles (e.g., “Java developer”).
Filter to profiles displaying email addresses.
Use scraping tools to extract emails, including:
- Instant Data Scraper
- Easy Scraper
- an email extractor browser extension
Manually open profile pages to recover embedded/visible emails where scraping may fail.

Concrete example metrics

Starting point example: 3,190 LinkedIn results labeled “Java developer” with emails shown.
Scraping outcome (with email extractor):
- 82 extracted via extension
- potentially ~100 total after accounting for embedded emails that scraping might miss

D) “Must-have / Nice-to-have” narrowing (search-string reduction strategy)

He recommends a narrowing loop:

Start with Must-have (e.g., core skill keyword + location/site).
Add Nice-to-have filters gradually to reduce volume to something manageable.

Example (Java Developer)

Initial scale: 10,400 items for “Java developers” (SlideShare-scale example)
Apply must-have/narrowing:
- reduce to ~1,500
Add nice-to-haves progressively:
- e.g., “J2E” / “Angular / SQL / C++ / Unix”
End goal:
- around ~84 results (described as “manageable” for outreach)

Playbook principle

Big → learn → small: start broad to see what appears, then tighten.

E) Experience-based requirements using graduation/dated anchors

For harder requirements like:

“10+ years experience”
“fintech/banking”
“Java”

He suggests using dated anchors found on profiles/resumes:

If a profile includes an entry like 2014, interpret it as likely graduation year or job year.
Then infer roughly 10 years of experience as of “2024”.
Combine with banking/fintech terms (e.g., “Banking” and related keywords).

Concrete example outcome

He points to cases where:
- one candidate shows 17 years of experience
- another appears consistent with 10+ after applying year + banking keywords

F) “Anchor words” for profile parsing (cross-platform)

Every profile/repo tends to include predictable category labels (“anchors”), such as:

present / contact / email / about
sections like: experience, education, skills, projects, volunteer
on GitHub-like pages: repos, commits, packages, stars, languages
on LinkedIn-like pages: “at” (commonly used for profile location context)

Operational tactic

Use anchor words to “tear apart” each site’s HTML/profile structure via search terms.

G) Hiring-manager-proofing the pipeline (avoid “dart throwing”)

He contrasts:

Dart thrower: random bulk resumes into outreach; may look similar but fail on contactability.
Targeted sourcer: ensures contactability (email/phone) and filters for fit, not just similarity.

H) Contact info extraction beyond LinkedIn (GitHub repo trick)

Some code-hosting platforms can contain emails not obvious on the profile.

Examples of methods mentioned

Tool: Email from GitHub
URL manipulation using repository commit pages:
- adding a suffix like “.patch” to reveal data (including email)

Actionable recommendation

Check code hosting artifacts (commits, patches, page source) for contact tokens not present in the public profile header.

I) Contacting using name + age estimation (approximation method)

When asked how to infer age:

He estimates age from profile timelines (education/employment dates), including military career progression examples.
He optionally references a tool that estimates approximate age from a picture, but suggests timeline-based estimation is often sufficient.

Example method (military timeline)

Assume entry age (e.g., ~18)
Use time-in-service requirements to reach a particular rank/role
He claims the inferred range was accurate “within two years.”

Organization/management implications (recruiting ops)

Recruiters often rely on shallow LinkedIn searches, repeatedly seeing the same top lists.
His approach expands sourcing to 350+ sites/tools (referencing large directories of additional platforms).
Goal: provide new candidate pools not already exhausted by hiring managers.

Operational outcome

More unique candidate sets → fewer duplicates from the recruiter “market pool”.

Metrics / KPIs explicitly mentioned (with examples)

No formal business KPIs like CAC/LTV are discussed. Instead, he cites sourcing volume and contactability-related numbers:

LinkedIn results limit: 1,000 results cap (on non-paid versions)
Email-enriched LinkedIn example:
- 3,190 candidates with email shown
Scraping example:
- 82 extracted emails via extension
- ~100 total at least after manual/profile verification
SlideShare scale:
- 10,400 Java developer keyword results
- narrowed to ~84 via Must/Nice-to-have filtering
Contactability notion:
- “you can get your 18% reply rate on instant messaging” (as rationale for improving contact readiness)

Concrete actionable recommendations (what to do next)

Build a synonym dictionary for every target skill (e.g., Java → J2SE/J2EE variants).
Use platform-specific anchor words:
- LinkedIn: rely on “present” and section anchors like “about/experience/education”
- GitHub/code hosts: use anchors like repos/commits/languages/stars/packages
Prefer image/PDF resume searches when text search yields low-quality results.
For outreach readiness:
- prioritize candidates with email addresses
- use scraping tools plus manual verification to recover embedded emails
Use a Must-have → Nice-to-have narrowing loop:
- aim for a workable shortlist (he targets ~100 or less)
Avoid “dart throwing” by ensuring both:
- likely skill fit
- likely contactability

Tools / product mentioned (supporting execution)

Source links.com
- Tiers: free (bronze/silver/platinum/gold) described
- Free covers about ~10% of capabilities vs paid (as stated)
- Includes a “search organizing” workflow so recruiters don’t repeatedly retype/copy-paste strings
Extensions/tools referenced:
- Instant Data Scraper
- Easy Scraper
- Email extractor extension
- “Email from GitHub”
- “FaceCheck in LinkedIn” (image/age lookup concept)
- Browser/Google advanced operators (Boolean + dorks)

(He also mentions a “Weebly search” on 30. weebly.com later, and a “Upr.com” reference, but the main operational focus is Source links.)

Presenters / sources (as mentioned)

Dean Da Costa — guest (“sourcing expert”; creator of the “Source/Sourcing Links” tool; provides the screenshare)
Dan — host (welcomes and interviews Dean)

The Sourcing Wizard: Dean Da Costa's Advanced IT Recruiting Strategies Screenshare

Key takeaways