How Founder-led sales can be done right? • Founder at Sellestial • 3rd-time Founder • Don't be a stranger, follow me and reach out

Is one of your main projects for 2026 cleaning up your HubSpot data?  Want to run AI in production?

If you answered yes to anything, keep on; otherwise, stop reading this.

👉 How we are different to tackle data hygiene with your internal team only:

1. Deep Data Cleaning & Enrichment Expertise
We're data cleaning and enrichment specialists who know this problem space and its traps inside out. By cleaning and enriching 100+ HubSpot portals, we've developed a deep understanding of edge cases that first-timers don't anticipate. This is not general HubSpot admin work; it's pattern recognition from seeing hundreds of messy databases.

2. Dedicated Bandwidth + Project Discipline
Data cleaning is important but rarely urgent, which means it gets perpetually deprioritized by internal teams. We provide dedicated focus with clear project boundaries that prevent the "cleaning to infinity" trap or arbitrary abandonment that internal teams tend to fall into.

3. Production-Grade AI Expertise for Data Maturity
We've built specialized AI infrastructure designed for accuracy and reliability. We are regularly deploying 7+ different AI models for different parts of the process. We tested and evaluated over 50 LLM models over the course of the last few years, and have a comprehensive understanding of their capabilities and caveats. 

The ROI for you: Your internal teams and HubSpot admins can focus on what they do best while we handle the extremely complex, time-intensive data cleanup and/or enrichment with the AI resources your team doesn't have access to.


👉 Want to talk about Sellestial?
Shoot me a DM, and we can schedule an introductory call to discuss data.

Klemen Hrovat

Want better HubSpot marketing email open rate?

Fix this.

Regularly review your marketing contacts:
- remove non-personal contacts
- remove disposable domains
- check for mx records of domains
- remove fake contacts
- check contact's employment status

Clean recipient lists  --> higher open rates

#EmailMarketing #HubSpot #marketinghub

Talks about HubSpot data + AI daily | HubSpot Community Champion

A customer asked us to help them with lead scoring in their HubSpot

saying "We have data in a separate system."

They're paying for a subscription to an industry-specific database. 
The data is supposed to help them score leads.

But the HubSpot integration?
Broken.
And the vendor has zero intention to fix it.

So they asked: "Can you help us get this information into HubSpot somehow?"

We built an AI agent instead.

It takes basic contact and company info from their HubSpot, searches for the specific information they need, and writes it back to the CRM automatically.

Here's where it gets interesting:
The agent finds more accurate information than their paid database.
It's more up-to-date.
And it has a higher success rate at finding what they need.

They're now questioning why they're paying for that subscription at all.

This isn't magic.
It's just what happens when you design data flows around what you actually need, not around what vendors decided to give you.

The real unlock?

When the right signals live in your HubSpot, lead scoring works.
Relevant lead prioritization becomes possible. 
Marketing is handing over good leads.
Sales stops wasting time.


Curious to hear from other RevOps and HubSpot admins.

What's a data source you're paying for that you suspect could be replaced or improved by something smarter?

Your reps are wasting 5+ hours per week chasing wrong accounts that:

- Are duplicates of accounts that they already reached out to
- Closed down companies
- Moved to another job 8 months ago
- Aren't actually a good fit

This isn't a process problem.

It's a data problem.

I saw that in a portal recently ... 9 duplicates of Coca Cola, missing subsidiaries, wrongly enriched company records ...  the list goes on up to 43 companies with something like Coca Cola as a company name.

Multiply that by every corporation in your HubSpot 🤯 

Before you hire more people or buy another shiny AI tool, audit your CRM. 

You will be surprised by the value hidden in the data you already have. 

What was the strangest duplicate situation you found in your HubSpot?

I got an idea to help HubSpot users understand their data AI readiness.

What could be improved in how I set it up in HubSpot?

I developed a form with 10 questions in HubSpot.

Each question is a new property with 3 possible answers.

Each of the possible answers has a different score (0, 1, or 3)

I didn't know how to weight specific values in properties to build the logic for how many points you get for each question natively in HubSpot. So I built a simple script in Sellestial that calculates the AI readiness score and writes it to a property in HubSpot.

I set up a workflow that, based on the calculated score, customizes the email with the score and some thoughts about their AI readiness and sends it via a HubSpot sequence.

HubSpot experts, what could I do better?

p.s. Want to get your AI readiness score? Linki in the comment

RevOps teams are accountable for outcomes they can't control.


- You own forecast accuracy but don't control pipeline creation.
- You own the tech stack but can't enforce process adoption.
- You're measured on conversion rates but have no authority over how leads get handed off.

That's not tools problem.
That's not a headcount problem.
That's a mandate problem.

What does RevOps actually own?

Have you set up the cap on marketing email frequency?

If you have already or plan to, be cautious.

This lets you cap the number of marketing emails each contact receives within a set period to prevent over-communication.

I keep seeing that challenge with HubSpot users I talk to.
Marketing sets up several/numerous/countless workflows. 
They automate marketing emails and lead nurturing.
But they don’t think about customer’s experience. 

A human got 2 emails right away, in the first 10 minutes after form submission.
Who knows what they’ll get today 🙂

By setting this up, you define a cap for regular marketing emails, workflow automated emails, and blog notifications. It does not apply to transactional emails, one-to-one emails, feedback surveys, or conversations inbox emails.

The cap is calculated per contact on a rolling time basis. If someone receives an email on Friday with a weekly cap of 1, they won't get another until the following Friday.


Think about these things before you turn this on.

1. You can’t guarantee that emails will be really sent to a contact. If you have some emails that you want to make sure you send, you can exempt specific emails (like critical announcements or promotions) from the cap by unchecking "Apply send frequency cap" in the email's send settings.

2. With that turned on, each email you send should stand on its own; it should not assume the person has received the previous email in the workflow.

When was the last time you reviewed your HubSpot contacts?

The 90-day data decay problem
Why your HubSpot Data Cleanup won't stick
and what to build instead


You just finished a cleanup project.
Reports finally make sense.
Properties standardized.
Duplicates merged.

Give it 90 days. 

It'll be broken again.

Why cleanups fail

We've audited over 100 HubSpot portals. The pattern is always the same: a team spends weeks (or months) cleaning their CRM. Three months later, they're back where they started, or worse.

The problem isn't the cleanup. It's the cleanup model.

One-off fixes are a lie.
They feel productive but decay within months. 
Nobody owns ongoing accuracy.

The only real solution is a system that run themselves.

We call this Perpetual DataOps. 

It's a framework for building HubSpot data infrastructure that stays clean, not for a quarter, but forever.

The goal of the Perpetual Data Framework: Stop fixing data. Configure a system that prevents data decay.

Three pillars

1) CLEAN DATA - Duplicates merged. Gaps filled. Properties standardized.
- <2% duplicate rate on contacts and companies (continuously maintained, not one-time)
- 95%+ fill rate on "must-have" properties
- Standardized formats (not enforced at entry, but standardized right afterwards)

2) RICH DATA - ICP scores. Buying signals. Lead prioritization.
- Contact + company enrichment running on every new record
- Waterfall enrichment from reliable sources
- ICP fit score configured and maintained
- Relevant signals captured and associated to the right records

3) UNIFIED DATA - Systems synced and unified.
- Field mappings explicit (no "sync everything and hope")
- Potential data overwrites defined with guardrails

The build plan

Days 0–10 | Audit & scope
Document your current data model
Identify your "must-have" fields per object (the 10–15 that actually matter)
Map where data enters: forms, imports, integrations, manual entry
Quantify the problem

Days 11–30 | Build the foundation
Deploy automated deduplication
Implement waterfall enrichment on contacts and companies
Configure processes to standardize field formats

Days 31–45 | Add intelligence
Build lead scoring based on firmographics + engagement signals
Configure lead prioritization logic
Set up routing rules based on clean, scored data

Days 46–60 | Activate & monitor
Turn on 24/7 operation
Document: how it works, when to intervene, how to extend


Stop cleaning your HubSpot. Start building infrastructure that keeps it clean.

The companies winning with AI, with ABM, with outbound at scale - they're not the ones with 100% perfect data. They're the ones with systems that keep data good enough, perpetually.

Ship the system. Let it run. Move on to strategy.

p.s. Want to know where your HubSpot data actually stands? We can audit your HubSpot to show you what's broken, what it's costing you, and what it takes to fix it permanently. No obligation to continue after the audit. DM or comment below.

Stop adding to the mess (before you start cleaning it up)
Everyone wants to clean their historical CRM data.
That's the wrong place to start.

The first quick win in data hygiene
STOP ADDING TO THE PROBLEM

Perform a thorough HubSpot data audit to understand what's broken, with a focus on understanding root causes, not just symptoms.

Start cleaning up your CRM after you understand:
- the current state of data
- what's causing the mess
- where data is coming into HubSpot
- what data is relevant and important

By knowing where you currently stand, you can design an implementation plan toward clean data. 

Do you plan to clean up your HubSpot in 2026? Where do you plan to start?

Unified Customer View: Pipedream or Realistic Goal?
Join me, Casey, and Klemen Hrovat for a live workshop moving through the Unified Customer View Playbook.

🎙️ New to streaming or looking to level up? Check out StreamYard and get $10 discount! 😍 https://lnkd.in/dAaAJzce

Had a great conversation with Karen Ng at INBOUND about what it actually takes to run AI in production.

The insight that stuck with me: "Evals has become a core skill that product managers need to understand."

AI is non-deterministic. 

Same prompt, different results. 

So how do you trust it?

You build guardrails:
→ Human in the loop
→ Evaluations at every level (prompts, outputs, actions)
→ Citations that trace back to source data

Watch the clip for Karen's full breakdown on how HubSpot approaches this.

"How many customers do you have?"

Asked a few of our customers when we started working together. 

None of them could answer it from their HubSpot.

Here are the reasons
Enrichment tools were overwriting properties. 
Duplicate companies were created. 
Churned customers were not marked accordingly.

Everyone and every system creates data. Nobody owns the mess.

So many CRM can't answer basic questions:

- Who are our active customers? Unclear. 
- Where's our biggest opportunity? Depends who you ask. 
- Which segment is growing? Check the spreadsheet backup.
- What do we report at BoD? Give me a week to prepare the slides.

You can't segment. Can't forecast. Can't prioritize.

I've watched RevOps spend hours manually building lists because they couldn't trust data. 

Seen execs commit to decisions based on numbers that were flat wrong.

All because nobody defined, implemented, and maintained data standards.

Want better insights? Fix data first.

And here's the part most teams skip: it can't be a one-time project.

Data decays.
People leave.
Processes drift.

Whatever you build needs to run itself, or you're back here in six months.

If you don't protect the foundation, nothing built on top will hold.

🎙️ New to streaming or looking to level up? Check out StreamYard and get $10 discount! 😍 https://lnkd.in/dAaAJzce

Roman Hipp

Building BetterContact | Helping early-stage B2B SaaS companies with insights & ready-to-use templates

Currently building BetterContact, a B2B SaaS that aggregates various data sources to find and verify any email and mobile phone number.

Growth Co-Founder

In 2024 alone, BetterContact lost $1.02M ARR to churn. In a fight to save the company, we turned the product into something users actually love. Monthly churn: 31.6% -> 8.4%.

From the beginning, we were solving a painful problem (bad email/phone enrichment quality), but our execution was shaky.

↳ Our contact data accuracy had gaps
↳ The enrichment execution was *super* slow
↳ The UI created friction (and users left quickly again)
↳ And we were missing a key piece: a proper list building engine

We knew that with >30% monthly churn rates, we couldn't build a company (without new customers, we’d lose almost everyone every 3 months).

So we threw everything we had at fixing the foundations:

↳ We ran 100's of A/Z tests on the data providers to refine our algorithm
↳ We invested a ton in our infrastructure to improve execution speed
↳ We onboarded +12 new data providers to close accuracy gaps
↳ We redesigned the entire UI to remove friction

And finally, we launched the first Waterfall Lead Finder - turning us into a full waterfall list building engine.

Super excited for what's to come in 2026, and how quickly we'll reach our target of sub-5% churn!

Building BetterContact | Helping B2B companies win outbound through more and better contact data

🎉 BetterContact officially partnered with Instantly.ai (we just went live!)

This is an especially exciting partnership for us because:
- we have a large joint customer base
- we ourselves are users for over 2 years now
- we’re getting along with the Instantly team very well

The big news for our users now:

You can now find emails from +20 premium data providers bundled in a single provider, directly inside one of the leading cold email sequencers.

The best thing:

They're verified and scored through 4 layers.
If they're invalid/have a low score, there will be 0 charge!

We've been in close contact with the Instantly team for a good while now, and really enjoy working together - kudos to the team for making it happen. 

Big things ahead in 2026!

Every "Find Contacts at Company" API is broken, because too often, the data is outdated. The person simply doesn’t work at that company anymore, or the job title just isn't the right one.

That’s why at BetterContact, we spent the last 8 months fixing it.

The result is the most advanced waterfall “Search People at Company” API endpoint, because:

1. It aggregates data from multiple sources (waterfall) - like we do with our enrichment waterfall + AI-powered web crawling & graph expansion

2. To ensure quality, AI Agents verify things like (in real-time):
- Does this person actually work there?
- Is this really their job title?
- Are they still based in that country?

The best thing: it doesn’t cost anything to source the leads, just the typical email/mobile phone enrichment rates apply (only charges if it’s triple-verified).

We’re opening up the endpoint to the people that comment, before we publicly release it soon.

The guide shows you how to easily set up the API.
(for Clay users, it also includes a template you can just use)

📩 Comment “Live” and I’ll send you the full setup step-by-step.
(And yes, you can use it today)

This is the allbound engine I used to bootstrap BetterContact to a 7-figure ARR SaaS - running 100% of sales and marketing myself with an inbound-led outbound motion.

𝐒𝐓𝐄𝐏 1: Gathering Intelligence

The best way I’ve found to truly understand my ICP’s pain points and how they design their workflows is by having real 1:1 conversations with them.

This gives me deep visibility into their exact use cases, while also helping me stay sharp on industry trends and competitor strategies.

This is what I’m using for that:

Call recordings & transcripts: Fathom.ai
Reddit keyword alarms: F5Bot
LinkedIn keyword alarms: Trigify.io

I automatically connect with ICP leads through HeyReach.io, and make sure to track their posts & activity through Breakcold.

𝐒𝐓𝐄𝐏 2: CREATE RELEVANT CONTENT

Those conversations directly fuel the ICP-relevant content I publish on LinkedIn.

When I have less skin in the game (i.e. fewer real customer and prospect conversations) the quality of my content drops fast.

That’s why STEP 1 underpins every step that comes after.


𝐒𝐓𝐄𝐏 3: TRACK ENGAGEMENT

I track profile views, post engagement, and other key signals using Teamfluence™.

This lets me spot people who might be interested in giving our solution a try.

𝐒𝐓𝐄𝐏 4: QUALIFY ENGAGEMENT

The engagers are being sent to Clay to qualify & enrich the leads.
(currently also evaluating tools like nRev, BitScale, Baseloop, and Genesy for that)

ZenRows for scraping, GPT o4-mini for ICP-fit scoring, and BuiltWith (occasionally) for technographics and integration angles.

For qualified leads (ICP Tier 1 & 2) I’m enriching the email address with BetterContact, and ICP Tier 1 leads also with mobile numbers.

GPT o4-mini to write personalized email copy.

𝐒𝐓𝐄𝐏 5: OUTREACH

My next step is reaching out to the different ICP levels.

The structure & cadence is the same (3 touches), but if there is no reply for ICP Tier 1’s, I try my best to call them. 

For that, I’m using BetterContact to find the mobile number, and Aircall to dial (depending on geography).

Something we recently added is Masterinbox.com to centralize all of our LinkedIn/email messaging (big helper there).

Outreach responses, from highly enthusiastic to openly annoyed, are a huge feedback loop for refining my understanding of the market and my ICP. 

Those insights, in turn, fuel better content, and the cycle repeats.

Would love any feedback on this process, or tool suggestions to make it more efficient 👇

Last week, a VP Sales at a Fortune 500 got personally offended after I told him we wouldn’t build a new feature into BetterContact for them to get onboarded.

Honestly, it was quite hard. Part of me wasn’t even sure if it was the right thing to do.

But looking back, I'm proud - it was the only right thing to do.

Because I knew: They’re currently not our ICP.

Sure, it’s tempting to say yes to a shiny logo like that.

But we've done this too many times in the past.

And ultimately, it always ends up in building the wrong product, for the wrong people, with the wrong priorities.

As our inbound requests increase significantly month-over-month, I'm sure there will be many more of those shiny temptations.

And every time, it's all abount reminding myself (albeit it's *really* hard):
Focus > FOMO

2025 has been a rollercoaster for BetterContact: $1.9M+ ARR in churn, 21% MoM growth, DDoS attacks, and much more. Here are some of our biggest fails and wins.

THE FAILS:

- $1.91M ARR churn (with >28% monthly subscriber churn spikes)
- Our best developer needed to leave due to personal reasons (we onboarded a new one though)
- Burned >$84k on features we couldn’t launch (things like a Sales Navigator extension)
- We still face occasional DDoS attacks which are very resource-intensive to defend against
- Major downtimes in Q1/Q2 that led to big customers churning

THE WINS:

- We kept innovating as first-movers (launched first waterfall lead finder)
- First in-person event co-hosted (Clay x BetterContact Christmas Party)
- Big accuracy boosts, especially with new partnerships like Trestle on the mobile phone side
- First time churn below 10% (currently at 8.2%!)
- Still very lean team with high EBITDA margins 
- We reached full data compliance with GDPR, CCPA, etc.
- Big new partnerships (Clay, Instantly, and a lot of other exciting ones I’ll share more about soon)
- No downtimes in Q4 (we struggled a lot with them in Q1/Q2)

In 2026, we’re doubling down on our mission: finding every discoverable email/mobile number, and thus enabling meaningful, mutually beneficial B2B connections.

Also, a big focus will be on improving & distributing our waterfall lead finder (we already get strong traction there).

Most importantly, we’ll do our best to stay laser-focused on building what our users actually want (and not chasing shiny objects or requests that won’t truly benefit them). 

Excited for the new year!

Six months ago, one of our biggest customer churned from BetterContact.

And honestly (kind of tough to say as the founder): I get why.

Back then:
↳ the UI was scrappy
↳ each enrichment took >5 minutes
↳ we had plenty of bugs and regular downtimes

Fast forward to today, we got this message 👇

“keep building, loving the product”

What makes this so great for us though isn’t the positive feedback itself.

It’s the person that it came from.

It's someone who left our platform when things weren’t good enough back then - and came back a month ago after we rethought and rebuilt the whole product.

Every improvement over the last six months came from moments like those early frustrations from our users. 

Still plenty to do, but moments like this remind me why the grind is worth it!

In the last 12 months, BetterContact grew primarily inbound, driving thousands of demos & signups by word of mouth. 3 weeks ago, we started cold calling again - and I forgot how crazy powerful it is.

The reason why we are picking up dialling again is:
1. AI makes everyone sound the same - but the phone is still where real human nuance can win
2. Our own data sourcing & enrichment is a competitive advantage

Up until now, our growth engine has been primarily inbound (word of mouth). 

But now we want to 10× what’s already working organically by layering cold calling back on top.

Below is a screenshot of our first campaign which went nuts.

65% connect rate (never seen that before - our big focus on mobile validation & scoring inside BetterContact seems to be paying off big time).

Based on those results, it would honestly be silly to *not* triple down on calling (we might put some open positions out there soon).

Are you currently picking up the phone? 👇

How to book a demo with billion-dollar CROs (this worked two timed for me in 2025):

→ build a data company
→ cold call them on their mobile
→ they reply a bit surprised: “I never get cold called - how did you find this number?”
→ they add: “…we actually need this for our sales team”
→ meeting booked

I never expected to get attention *this* way for BetterContact, but not complaining!

“We just launched the most accurate lead database”. And still, it thinks I’m working at my cold company. This is the reason why many outbound set ups are broken (and what we did about it) 

The big problem with these databases is the following:

They only periodically scrape platforms like LinkedIn, and as they need to scrape 10’s of millions of profiles regularly, there is a significant data loss.

Scraping timeouts & omitting data points on purpose to be more efficient (LinkedIn gets better and better at preventing that at scale).

And if someone changed jobs recently, tools like Apollo still have their old info, rendering these prospects useless for your list building.

That’s the reason why 8 months ago, we wanted to fix that at BetterContact (primarily because it was the most requested feature by far).

Based on our +250 beta testers (just quoting some users here):

“It worked very well, especially compared to what the legacy databases can do”
“I’m annoyed with Apollo’s shi*** data. This is a game-changer”
“Definitely the freshest people search API I used so far”

Because this is what our waterfall API does differently:

We waterfall databases in order to solve the issue of the coverage.

And to solve for quality/accuracy, we use AI Agents to crawl the web in order to verify that the person *actually* works at that company/has that job title/lives in that country.

We already have it realeased in the app, and next week we’ll release the API publicly. Keen on every feedback!

taplio

The Battle

Klemen Hrovat

Roman Hipp

Followers

Posts

Engagements

Likes

Performance ScoreKlemen Hrovat's LinkedIn Performance

Performance ScoreRoman Hipp's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Klemen Hrovat

Roman Hipp

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);}Klemen Hrovat's LinkedIn Performance

Performance ScoreRoman Hipp's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreKlemen Hrovat's LinkedIn Performance