What I Learned Building a Data Tool

Emily Goh · 2026-04-02

The moment I had enough.

I've been a data leader and a marketing leader. Mostly marketing, but data is where I've lived and breathed for the past 15 years. Growth. Demand gen. Pipeline numbers, MQLs, cost-per-lead, and trying to explain to the CEO why we need more budget for paid ads.

My job is to feed the sales team with enough pipeline, defend every dollar I spend on campaigns, and walk into board meetings with a story that holds up. Revenue. Attribution. What's working. What's not.

And here's the thing nobody talks about: all of that depends entirely on trusting your data.

I had a spreadsheet. My team manually entered costs every time we got a receipt, an invoice, a contract renewal. Finance would then come to me at month-end with actuals vs. budget questions. My spreadsheet never matched their system. Never. Not once.

So I'd spend hours reverse-engineering. Going back through invoices. Cross-referencing what was in HubSpot vs. what was in the GL. Trying to reconcile where I'd spent $40k that wasn't showing up in the right bucket.

And every time finance came back with a question, a small part of my credibility went with it.

The credibility problem.

When you're a marketing leader, your data is your defense.

You're spending company money. On ads. On events. On headcount. And someone is always going to question it. Finance. The board. Sales, when pipeline doesn't convert.

If your data doesn't hold up, if someone finds a gap before you do, that's a career moment. Not the good kind.

So I did what every marketing leader does: I pulled everything manually. Built my own source of truth in a spreadsheet. Compared systems line by line. Spent Sunday nights before QBRs making sure the story was airtight.

Not because I loved spreadsheets. Because I had no other way to trust my numbers.

And then I'd present the data, MQLs, CPL, pipeline contribution, and someone would point at a gap I hadn't caught. I'd have to go back, dig in, come back the next day with an explanation.

Hours I didn't have. Credibility I couldn't afford to keep losing.

The moment I decided to build this.

I'd had enough with the gut feeling.

There's a version of marketing leadership where you make decisions from instinct and then reverse-engineer the data to support them. I'd been doing that, not because I wanted to, but because I couldn't trust the data fast enough to lead with it.

I was comparing systems to find missing leads. Running deduplication manually. Checking if the same contact existed in HubSpot three times under different email formats.

The insight that pushed me over the edge: a tech person can build anything. But the person who has gone through the trenches, who knows what it feels like to defend a $200k campaign budget to a skeptical CFO with a broken spreadsheet, that person can build something the engineers never would.

I'd been that person for years. So I started building.

What I got wrong in V1.

I thought people wanted visualizations.

Dashboards. Charts. A beautiful story about pipeline velocity and campaign attribution. I thought the problem was that marketing leaders and ops teams couldn't see their data clearly enough.

I was wrong.

The problem isn't that the data isn't visual. The problem is that the data is bad.

You can have the most beautiful chart in the world. Wrong data underneath it. Wrong decisions made from it. And sometimes it's innocent people whose heads roll because leadership made the call on numbers that were never right to begin with.

Visualization doesn't mean anything when the foundation is rotten. I had to rebuild the whole premise.

What actually matters.

Flag the bad data first.

Before you tell a story. Before you build a chart. Before you walk into a board meeting. You need to know: is this data trustworthy?

That's what Gohvio does. And then we layer Lumi, our AI layer, on top of that human check, the way any good AI usage should work: not replacing judgment, but covering ground faster. Flagging what a human might miss. Surfacing the gap before you step in it.

It's not about pretty. It's about confident.

What surprised me.

People don't trust their data, but they've normalized it. They feel the uncertainty. They just don't know what to call it or how to fix it, so they live with it and hope for the best.

Process and data quality are essential, but people treat them like they're nice to have. Until something breaks. Then it's an all-hands emergency. The teams who invest nothing in data quality are the ones paying the highest price for it.

The problem is universal. Doesn't matter if you're in marketing, sales, finance, or ops. Doesn't matter if you're on HubSpot or Salesforce, NetSuite or QuickBooks. The problems are identical: duplicates, stale records, missing data, systems that don't talk to each other.

Knowing about the bad data is already valuable, even before you fix it. You can prepare. You can caveat. You can go back to finance before they come to you. That shift, from reactive to proactive, changes how people experience you in a room.

Why I'm doing this.

I built Gohvio because I know what it's like to stand up in a board meeting and quietly hope no one finds the gap you didn't have time to check.

I know what it's like to have finance walk in with a question that takes 48 hours to answer because the answer lives across three systems that don't agree.

I know what it's like to lose credibility, not because you were wrong, but because you didn't know you were wrong.

Marketing leaders, ops leaders, finance leaders, they have tools to store data. They need a tool to know whether it's worth betting on.

That's why Gohvio exists.

We're in early access. And the most common thing I hear from early users is: "I can't believe I was making decisions without this."

Neither can I. I just needed to build it first.