Hey friend, how’s it going? 🌻
Marketing folks, especially content and SEO people, have a bad habit in general: they focus on one tool, a maximum of five data points, and that’s all.
However, as you transition to full-stack marketing, you will know that you need to dive deeper. There are five places I go regularly to understand why something's working, and Search Console and GA are only two of them.
Let’s get into it 👇
📌 TL;DR
Each data source shows you a fragment, not the full picture, treat it that way
Search Console and GA answer “what happened,” not “why”
Your CRM only sees someone after they’ve already decided you’re worth their time
LinkedIn comments carry objections weeks before they’d ever show up in a CRM
Call transcripts have the actual words your buyers use, most teams never open them
Heatmaps show you the exact sentence where you lost the reader
#1 Search Console + GA: the “what,” not the “why”
Does a traffic spike ever actually tell you why it happened? No. It tells you it happened. You still have to go find out why.
Two things I’ve learned the hard way here:
A wall of “direct” traffic in GA4 is rarely people typing your URL from memory. It’s usually missing UTM tags. So even your “what” can be broken before you’ve gotten anywhere near “why.”
Rank and impressions going up mean nothing if engagement on the page doesn’t keep pace. I’ve had pages jump five positions and lose scroll depth in the same month. GSC would’ve told you that’s a win. It wasn’t.
What I actually do: stop reading rank in isolation. Pull it against scroll depth and time on page for the same URL, same date range. If rank’s up and engagement’s flat or down, the query brought the wrong intent, or the page didn’t deliver on what the query promised.
AI angle: I run Ahrefs data straight through Claude via MCP now, instead of exporting it into a spreadsheet I’d open twice and forget. One prompt to cross-reference top queries against engagement and flag the pages where rank is strong but engagement is weak.
#2 CRM + customer journey: sees people only after they’ve already decided
By the time someone’s in your CRM, the actual decision-making has already happened somewhere your CRM has no visibility into.
There’s a pattern that shows up constantly in attribution work: a channel looks weak on last-click, but strong on assists. Dig in, and you’ll often find someone saw the touchpoint, didn’t act, then searched for the brand directly three weeks later. Your CRM logs that as “organic search.” It wasn’t. It was the channel you almost wrote off.
What I actually do: pull closed-won deals and trace back which content actually touched them before the demo, not what the attribution model assumes touched them. Those two lists rarely match, and the gap is usually where the real story is.
AI angle: If your CRM offers a connector or MCP, use it. Feed closed-won deals and your content calendar into Claude, and ask it to surface which pieces recur in the touchpoint history before conversion. The manual version of this is used to eat in the afternoon.
#3 LinkedIn: the objection shows up here first, weeks before it hits a form
This is the one source on the list that’s giving you the pushback before anyone’s even considering buying.
Someone comments, “Does this work for teams under 10?” Someone says, “We tried something like this, and it didn’t stick.” None of that reaches your CRM. Most of it never reaches your CRM, even after they buy, because by then it’s been reduced to a form field with none of the original language left.
What I actually do: keep a running note of the exact phrases people use, not my summarized version of what I think they meant. Their words, not my translation.
AI angle: less a tool, more a weekly habit. I paste a batch of the week’s comments and DMs into Claude and ask it to group by theme, nothing more, just show me what’s actually repeating. Catches patterns I’d have smoothed over on my own.
#4 Call transcripts: the most accurate source, and the one nobody opens
Guess which of these five gets used the least relative to how useful it actually is. This one. Nobody has forty minutes to sit through a sales call.
But this is where the real vocabulary gap lives. Marketing says, “Optimize your conversion funnel.” The prospect says, “We’re not getting enough leads.” Marketing says “omnichannel engagement strategy.” The prospect asks, “Should we even be on TikTok?” That gap is why a headline can be technically accurate and still not land; it’s written in the wrong language.
What I actually do: pull the last fifteen or twenty discovery calls, tag the objections and phrases that keep repeating, and use those exact words in the next headline instead of “professionalizing” them into something that sounds more polished and says less.
AI angle: This is a good candidate for a saved skill rather than a one-off prompt. Set up a repeatable Claude workflow that takes a batch of transcripts and consistently pulls the same three things: recurring pain points, objections, and exact phrases. Same format every run, so you can compare this month’s patterns against last month’s instead of starting from scratch each time.
#5 Heatmaps: where the sentence actually loses the reader
Suppose a content team found that almost 40% of readers dropped off right after the first CTA in their blog posts. Not at the end. Not halfway through. Right after the first ask.
That changed how they wrote intros. Shorter, a table of contents so people could skip to what they needed, and everything before that first CTA treated as the only real estate that mattered.
One caveat, though, low scroll depth isn’t always a bad-writing problem. Could be a slow-loading page, a mismatch between what someone searched for and what they landed on, or a layout that quietly discourages scrolling. Rule those out before you blame the paragraph.
What I actually do: pull the heatmaps for the best- and worst-performing posts side by side. Find the exact line where attention drops, and check whether that line earned the scroll or assumed it.
AI angle: heatmap tools don’t talk to Claude directly, so this one stays semi-manual. Export the scroll percentages, paste them alongside the actual paragraph text at each drop-off point, and ask Claude whether the writing likely lost people there or whether a layout issue is the more probable cause. Second pair of eyes on your own instinct, not a replacement for it.
📚 Check out my digital products & services purchased and loved by 250+ writers and marketers
AMA: SaaS Content Marketing | 45 Minutes: Strategy, writing, analytics, conversion, SEO, AEO, nothing is off limits. Share your website or portfolio in advance, and I’ll come prepared with specific answers rather than generic advice.
🎁 BTW, it’s my birthday today, and there’s a special 44% discount until tomorrow!
BOFU Content E-book: Get my updated BOFU Content e-book that covers three new chapters on BOFU content beyond alternatives and comparisons, BOFU attribution, and AI workflows for BOFU content.
AI+Human Content Playbook: For Content Marketers only. Essentially, for those under pressure to produce a large volume of content, it limits their ability to learn and experiment with other aspects of content marketing.
💡New update: A Claude skill.md file to accelerate your writing workflow! No price change.
Say Hi on LinkedIn 👋
Onwards and upwards
Sreyashi