LinkedIn Profile Optimization: A 90-Minute Rebuild

Each field does one job. Write it for the stranger, not for people who already know you.

Two jobs, in order. Being found. Then being believed. Most profiles pass the first and fail the second: a title, a company, and duties copied off an appointment letter.

This page covers the personal profile only. LinkedIn as a channel, with its cadence and formats, sits in our LinkedIn guide. Having something worth saying is a separate discipline, covered in our thought leadership post. Your company page is a different asset with different rules again.

What a LinkedIn profile is actually for

A profile is a landing page whose template you do not control. Its job is to turn a stranger's ten-second scan into a reply, a shortlist or a meeting.

LinkedIn's Pressroom statistics page puts the network at 1.3 billion members, with 410 million-plus across Asia Pacific (accessed 24 September 2026). It publishes no India-only figure, so treat every "LinkedIn India has X crore users" claim you read as unsourced until somebody shows you the page.

One running example throughout, entirely hypothetical: Tarakite Labs, a Pune B2B SaaS company with a founder (Rohit), a senior performance marketer looking to move (Ankita) and an account executive (Faiz).

How LinkedIn search finds you now

LinkedIn has moved search from matching your words to matching your meaning, which quietly kills the old advice about repeating one keyword in every field.

Its own engineers set out both methods in a paper on LinkedIn's content search engine (Yang et al., LinkedIn Corporation). Term-based retrieval "selects candidate posts that contain the exact same keywords as in the query". Embedding-based retrieval "uses an AI model to select candidate posts" by comparing meaning instead. Their reason for the shift: "Members are now developing the habit of issuing long and complex natural language queries."

That paper covers post search. The people-search equivalent is "Semantic Search at LinkedIn" (Borisyuk et al., submitted 7 February 2026), which covers "AI Job Search and AI People Search" and combines "an LLM relevance judge, embedding-based retrieval, and a compact Small Language Model".

The product shipped on 13 November 2025, first to Premium users in the US (TechCrunch). Product director Rohan Rajiv, on what it replaced: "With lexical search, you have to know the exact title of the person, or you need to wrestle with filters."

So a recruiter searching "someone who has run paid acquisition for Indian B2B SaaS" does not need those words on Ankita's profile. Repeating a phrase across eight fields does not help a system built on embeddings. One clear sentence does.

Which profile fields do which job

Sort every field into one of three jobs: findability, credibility, conversion. Optimising a credibility field for search is the most common mistake on this page.

On character limits: LinkedIn's Help Centre blocks automated fetchers, so we could not verify one field limit from LinkedIn itself. Every "220-character headline, 2,600-character About" figure online traces back to third-party blogs. Draft it, paste it in, and let the counter tell you.

Headline: five rewrites from bad to good

The headline is the only field that travels with you everywhere: search results, comment threads, connection requests, notification emails. Most people spend it restating their job title. Five patterns we see constantly, with what we would write instead (hypothetical people, illustrative wording):

• "Digital Marketing Professional" → "Paid acquisition for Indian B2B SaaS · Google & LinkedIn Ads · Pune". A category with no object becomes a job with a customer.
• "Aspiring Data Analyst | Passionate about AI | Lifelong Learner" → "Data analyst · SQL, Python, Power BI · Built the churn dashboard at a 40-person fintech".
• "Founder & CEO at Tarakite Labs" → "Founder, Tarakite Labs · Warranty software for Indian appliance brands". The title already appears in the Experience block below.
• "Helping brands unlock their true potential 🚀" → "Helping D2C brands fix returns before they scale ads". A claim with no noun gets a problem attached.
• "MBA | B.Tech | Ex-Deloitte | Open to Opportunities" → "Supply-chain analyst · Ex-Deloitte · Looking for ops roles in Mumbai". Credentials are past tense.

Our view (opinion, not data). For Indian B2B the formula that holds up is [what you do] for [who] · [one proof point] · [city]. Three reasons, from watching these profiles rather than from a study. Indian B2B titles are inflated, so a title separates nobody while the vertical does. Buyers here filter hard on location for anything involving implementation, and "India" is not a location. And a proof point, whether a named tool, a defensible number or a recognisable former employer, survives being skimmed where an adjective does not.

Applied to Ankita: "Performance marketing for B2B SaaS · ₹40L+ annual ad spend across Google and LinkedIn · Pune". Use a number only if you would be happy to be asked about it on a call.

The About section is a positioning document

Write About as four short paragraphs answering four questions in order: who you help, what problem you solve, what proof exists, and what happens next. Not a biography.

Social Media Today reported on 5 February 2024 that LinkedIn was moving About back to the top of profiles. We cannot verify its position in your current layout, so check. Either way it is the first block of real prose anyone reads.

• Line one, no run-up. "I run paid acquisition for B2B SaaS companies selling into Indian mid-market." Not "With over 8 years of experience across a variety of domains…"
• The problem you keep meeting. One sentence your reader recognises. For Ankita: pipeline that looks healthy at MQL and dies at demo.
• Evidence, not adjectives. Two or three specifics: a channel you rebuilt, a number, a named tool, a category you know cold.
• A next step. "Open to senior performance roles in Pune or remote, email is fastest." A profile without an instruction gets admired and closed.

First person. Third person about yourself reads like a conference bio from 2014.

Experience entries: outcomes, not duties

Two to four lines per role, each describing something that changed because you were there. Duties describe the job; outcomes describe you doing it.

Our test: could your replacement have written this line on day one, before doing anything? If yes, it is dead weight.

• Duty: "Responsible for managing Google Ads and LinkedIn Ads campaigns for the company."
• Outcome: "Rebuilt the Google Ads account around three problem-led search themes and moved demo requests from branded to non-branded traffic within two quarters."

Give the last two roles this treatment and leave the rest as a line each. Nobody reads role five. Where you cannot share numbers, describe the change rather than inventing a percentage: a system you built, a process that outlived you, a channel that did not exist before.

This is where the profile stops being a personal exercise. When we audit a client's presence at Digital Hangover, the founder's and the sales team's profiles get reviewed alongside the company page, because in B2B that is where the conversations happen.

Skills, endorsements and recommendations, honestly

Skills are a structured field LinkedIn genuinely computes on. Endorsements are close to noise. Recommendations do nothing for search and a great deal for belief.

The evidence for skills comes from LinkedIn itself. Its Skills on the Rise 2026 report (24 February 2026) states the method openly: it measures "growth of a given skill being added to member profiles" alongside skills held by members hired in the past year. LinkedIn's Pressroom lists 42,000-plus skills in the taxonomy. A controlled vocabulary, which is why five accurate entries beat fifty scattered ones.

That report named five skill stacks rising fastest in India: AI and automation, data and analytics, IT and cybersecurity, business and growth, and people and leadership (Storyboard18 on the India data, 24 February 2026). If you are deciding what to learn rather than what to list, our post on digital marketing skills covers that.

Endorsements: we found no LinkedIn documentation saying endorsement counts affect search ranking. Every article claiming they do asserts it without a source. Skip the reciprocal rituals.

Recommendations: ask with the specifics supplied. "Would you write two lines about the warranty rollout, especially cutting onboarding from six weeks to nine days?" produces something usable. "Could you recommend me?" produces "Great to work with, highly recommended."

Same goes for the 21× profile-views claim and the Open To Work statistics that circulate every January. We could not trace any of them to a live, dated LinkedIn source this week. Add a photo because a faceless profile is not believed. That reason is enough.

Featured, and the surfaces that used to need Creator Mode

Creator Mode no longer exists. LinkedIn removed the profile hashtag field in February 2024 and the toggle itself the month after, saying the features behind it were being opened to everyone.

LinkedIn's statement at the time, quoted by Social Media Today on 5 February 2024: "In the coming months we'll be opening up additional creation tool access and deeper analytics to all members… and removing the need to toggle creator mode on and off." The follow button, the Featured section and the deeper analytics stayed.

So a guide telling you to switch on Creator Mode to reach Featured is describing a product retired two and a half years ago. We cannot verify how Featured behaves today against LinkedIn documentation, for the robots reason above, so open your own editor before planning around it.

One first-party clue about what LinkedIn treats as structured data: its Full Profile API field reference (Microsoft Learn, updated 6 March 2026) lists headline, summary, positions, skills, certifications, courses, publications, patents, projects, honors and vanityName. That access level is closed to developers now, but the list is instructive.

Founders, job seekers and salespeople need different profiles

Same fields, different weighting. Pick which of the three you are this quarter, because a profile trying to be all three is none of them.

• Founder (Rohit). The profile sells the company before the website gets a chance. Headline carries what the company does, not the title. Featured carries one customer story and one way to book time.
• Job seeker (Ankita). Optimise for the recruiter filter and the shortlist scan: exact role titles in Experience, accurate city, skills matched to the postings you are applying to, and an About that ends with what you want.
• Salesperson (Faiz). Every prospect checks the profile between the email and the reply, so the banner and About should read like a reason to take the meeting. Outbound mechanics are a separate subject and sit in our LinkedIn lead generation guide.

The 90-minute profile rebuild

One sitting, phone off, no design work. The headline and About get half the block because they do most of the work.

1. 0–10 min · Collect raw material. In a plain text file, write five things you did in the last two years that changed something, each with a number or a name attached. Everything below is edited from this list.
2. 10–20 min · Photo and banner. A recent photo where your face fills the frame, and a banner carrying one line of positioning, checked on a phone rather than a laptop.
3. 20–35 min · Headline. Write five versions using the formula above. Read each aloud. Keep the one that still makes sense to someone outside your industry.
4. 35–55 min · About. Four paragraphs: who you help, the problem, the evidence, the next step. Delete any sentence that could sit unchanged on somebody else's profile.
5. 55–70 min · Your last two Experience entries. Two to four outcome lines each, taken from step one. Older roles get a single line.
6. 70–78 min · Skills. Cut to five to ten from LinkedIn's list, matched to the roles or deals you want. Pin the three that matter most.
7. 78–85 min · Featured. Three items: your strongest proof, your best-performing post, and one link that lets somebody act.
8. 85–90 min · Housekeeping and two asks. Set a custom profile URL, fix the location to your city, check the contact email works. Then message two people with a recommendation request that names the project.

What to do weekly once it is set up

Twenty minutes a week, none of it spent editing the profile again. The profile is the destination; the weekly work sends people to it.

• Comment on five posts in your market, specifically enough that the comment could not be pasted under any other update.
• Publish one thing you would say out loud in a meeting. One post a week beats five in a burst and a month of silence.
• Log where inbound came from. A note each time someone says "I saw your profile" tells you in eight weeks whether any of this worked.

Rewriting the headline a fourth time is not work. It feels like work, which is why people do it instead of posting. Cadence and formats belong in our LinkedIn marketing post.