David Ly Khim

I'm a Cambodian-American entrepreneur and growth practitioner. Currently building Omniscient Digital, an organic growth agency helping B2B software companies turn SEO, LLMs, and content into revenue growth channels.

Previously, I was Head of Growth at People.ai where I led the implementation of a product-led growth motion for the new freemium products, PeopleGlass and ClosePlan, overseeing product strategy, go-to-market, and customer acquisition initiatives.

Before that, I spent 6 years working on HubSpot's freemium business through multiple cross-functional projects including: driving the implementation of referral loops across the free product, bottom-of-funnel content strategy, and business development.

One key project I worked on was developing and executing a strategy to acquire users through marketplaces and integrations, which required collaboration across marketing, business development, and multiple product teams. This resulted in spinning up a new Growth Acquisition Team in Dublin. We revived and grew adoption of (a previously sunset) HubSpot WordPress integration from 8k to 110k+ installs (adopted by 22% of the customer base), drove 1000s of product signups a month, and influenced $3.6M+ in ARR. 

Skills: product management, growth experimentation, growth marketing, data analysis, user research, user acquisition, project management, SEO, content strategy, partnerships

Personally, born and raised in Long Beach, CA, now based in Boston. Into weight lifting, running, Muay Thai, reading, and whiskey. Not all at the same time.

Co-Founder, Head of Strategy & Operations

There was a period in SEO where you could win without ever talking to a customer.

Pull some keywords from ahrefs.
Write and publish some optimized pages.
Rank, bring in traffic.
Profit (maybe).

That era has been over but you'd be surprised at how many teams are still trying to run that playbook.

The teams winning now have a deep understanding of their customers' psychology

They know their customers’ pain points, ambitions, and the language they use. And it shows up in the content.

One company we work with sells financial planning software.

They’re frequently cited in ChatGPT.

It's not because they publish a bunch of listicles.

Instead, they've become a resource to help financial planners how to become better at what they do and to build their book of business. They share methodology. They have a point of view. They build topical authority.

So their brand get referenced even when someone isn’t asking ChatGPT for software recommendations.

They appear early and often in the customer journey.

If you’re not deeply embedded in your customer’s world and speaking to what they actually care about, you might be publishing content into the void.

Co-founder at Omniscient Digital | Ex-Growth @ HubSpot, People.ai

LLM visibility is having its vanity metric moment.

We just swapped keywords for prompts.

When measurement is unclear, marketers default to what’s easiest to count.

That’s the streetlight effect.

If you drop your keys at night but the only light is under a streetlamp, you search under the streetlamp. Not because the keys are there, but because the area with the most visibility.

In GEO, the streetlamps are:
- Category prompts (“best CRM for startups”)
- Trackable clicks from ChatGPT referrals
- Rankings for a small set of queries

Those are fine.

But if you only optimize for what you can measure, you miss what’s actually happening:

Whether we like it or not, this is the current nature of LLMs, where much of the influence they drive is upstream of measurable behavior.

It’s the same dynamic as zero-click LinkedIn content.

You read a post.
You remember the brand.
You search later.

Attribution says “organic.” Reality is “influenced by a LinkedIn post many weeks ago”

So what do we do?

- Track the streetlamp metrics
- But build a strategy for upstream influence
- And accept that all decisions include some level of uncertainty. Properly collected and analyzed data can reduce that uncertainty, but never eliminate it.

A month ago, G2 agreed to acquire Capterra, Software Advice, and GetApp.
Why does this matter for AI visibility?

One company will now control four of the most authoritative review domains in B2B software.

When a buyer asks ChatGPT "best CRM for a 200-person company," the sources cited in that answer shape the shortlist.

Review sites are heavily cited at this stage of the customer journey, the exact moment they're narrowing down their options.

Cate analyzed 25,000+ AI citations across ChatGPT, Perplexity, AI Overview, AI Mode, and Gemini and found something interest:

When you model the four domains together, G2's share of AI citations in bottom-of-funnel prompts jumps 76%.

They move from the 4th most-cited source to #2 (behind only Reddit).

In social proof queries specifically (customer reviews, case studies, testimonials), their combined share would give them a 93% citation lead over the next closest domain.

This is a consolidation play for the AI layer, not just the review layer.

For B2B software companies, the implication is straightforward:

The surfaces that influence AI answers are consolidating. If your review presence is weak on G2 and its new properties, you're losing influence at the point in the buyer journey that matters most.

Link to the full analysis in the comments.

I help companies measure and take action on their marketing data.

We both know that being online takes a lot of work. I help eliminate the "Guess Work". 

Using Google Marketing Cloud we are able to predictably increase MRR while measuring each channel's step in the marketing journey.  

We take all of your Data from your CRM, Google Analytics, and cart platform, then turn it into a  real-time data dashboard for you to take action.

What most digital businesses need is someone to provide reliable consistent data to take action. You have a team and a plan, but want to know "Is it working?" & "Where can we cut the fat?".

Head to Mediauthentic.com and apply to work with us. 

THE COMPANIES THAT WE WORK WITH! 
✅ Care about your customers! 
✅ Want to grow with as little wasted marketing efforts as possible.  

WHAT WE DO! 
We offer custom solutions that help you: 

☑︎Grow your MRR (Monthly Recurring Revenue).

☑︎Increase the velocity of return (Aware -> Revenue). 

☑︎Measure your current system efficiency. 

☑︎Pivot into more reliable traffic solutions.

☑︎Alway improve your measurement systems.


WHAT OUR CLIENTS ARE SAYING
“Accountable, knowledge, and able to deliver on time. You won’t regret working with JJ.” Russell Kim, DDS - Windward Smiles

“So amazing to work with! it seems like they get stuff done 10x faster then I ever would have thought!” Shilhi Seibel - Passion Roots 

“Mediauthentic helped me create an online medical education platform. They helped me feel proficient with the technology applications I needed for my projects. What I appreciated most was that JJ took the stress out of the process (actually made it fun!). He was easy to interact with and kept me on track to reach my goals” - Darin Gould, MD


JJ Reynolds

Imagine this world -
Product & Marketing analytics in the same location 
&  
engineering teams aren't bogged down from marketing asking for " one more thing"... 

Check this out 
https://lnkd.in/g7AR8zES 

PostHog endpoints fueling a web analytics custom UI. 

Time to action is getting shorter & shorter for marketing & product teams.

Helping marketers make better use of their data | custom tracking, visualization & forecasting

5 tools to do one job.

That's the average data stack in B2B SaaS today.

PostHog bet on convergence. And it shows:

➡️ One single event schema for everything

➡️ Analytics and routing from the same source

➡️ ~50k MAU / 5M events: $615/mo vs $3k-$8k+ with Amplitude + Segment

➡️ No connection fees, no license Tetris

Where does it fall short? Segment is still better with 50+ destinations and complex transformations.

But for most teams: fewer tools = less money = one single source of truth.

Is your stack still fighting itself? 👀

No alternative text description for this image

AI makes data analysis and activation way faster.

But if your data foundation is broken, you're just automating bad decisions faster.

And most companies have a broken foundation.

Before you touch AI in your stack, you need to have this:

➡️ Governed event taxonomy

page_viewed vs Page Viewed vs pageView across three SDKs = data that looks complete but can't be joined.


➡️ Identity resolution across sources

Without a layer that unifies anonymous ID, email, and domain, your AI is analyzing ghost users counted three times.

➡️ Data contracts and schema enforcement

When a dev renames a property, your pipeline should break loudly. Not silently pass malformed data for weeks.

➡️ Modeled entities in your warehouse

AI querying raw events is going to hallucinate joins and miss business logic every single time.

➡️ Quality monitoring

Anomaly detection, freshness checks, drift alerts. If you don't know your data is broken within hours, nothing downstream is reliable.

Get this right and AI becomes a multiplier.

Skip it and you're just automating bad decisions faster.

Thanks for all the DMs with suggestions! 
Now you can easily filter all metrics & dimensions in real-time :) 

Using the PostHog endpoint. 

Also;
Pricing is unknown 
Response time is debatable 
Coolness is great!

A lot of teams aren’t staying on GA4 because it’s working especially well. They’re staying because ripping it out feels like a pain, and nobody wants to be the person who creates even more tracking chaos.

Meanwhile the chaos is already there.

Purchase events don’t quite line up with actual revenue. One country domain tracks one way, another breaks in a totally different way. Someone tries to fix it server-side and suddenly one market looks cleaner and another one looks worse. And after a while nobody really says it out loud, but nobody fully trusts the numbers either.

So the team just... lives with it.

Keeps patching. Keeps adjusting. Keeps looking at dashboards with that quiet feeling of “this is probably directionally useful, I guess.”

That’s usually the point where I think the GA4 vs PostHog conversation gets framed wrong.

PostHog isn’t just “another analytics tool.”

It changes the shape of the setup.

With GA4, you’re mostly working inside one system and trying to make your needs fit around it.

With PostHog, the data can move where you need it to move.

Sessions, personas, events. Into BigQuery. Into Slack. Into HubSpot. Into feature flags. Into your warehouse. Into the rest of the stack you actually use.

So it’s not really about “starting over.”

It’s more like finally setting things up in a way that matches how your team already works.

And honestly, for most teams, the real blocker isn’t love for GA4.

It’s fear of the mess of switching.

Which is fair.

But staying in a system nobody fully trusts is its own kind of mess too.

What’s actually keeping your team on GA4?

Actionable Insights come at the intersection of expectation x outcome x ability to act. 

Useless scenarios: 
You do not have an expectation & know the outcome
You do not have the ability to act & have an expecation 
You know the outcome & have the ability to act, but there was not expectation.

There are three games to win. 

Game 1: Own the underlying data at a cost everyone supports 

Game 2:  Own the interaction of the data 

Game 3: Own the opinion on how to action the data. 

Each game increases in complexity and downside if done wrong.

taplio

The Battle

David Ly Khim

JJ Reynolds

Followers

Posts

Engagements

Likes

Performance ScoreDavid Ly Khim's LinkedIn Performance

Performance ScoreJJ Reynolds's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

David Ly Khim

JJ Reynolds

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}David Ly Khim's LinkedIn Performance

Performance ScoreJJ Reynolds's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreDavid Ly Khim's LinkedIn Performance