# Top 10 in Tech - What to know for Week ending June 26, 2026

Published: 26 June 2026
Canonical: https://www.top10in.tech/posts/top-10-in-tech-what-to-know-for-week-ending-june-26-2026

## 1. SaaS METRIC OF THE WEEK

CAC PAYBACK: The 'payback' period is the nuance of why we measure CAC. How long until we break even? Benchmark-wise, the negative trough is way longer than you think, so take a seat! New B2B customers, on **_average_**, [take 2 years and 2 months](https://www.dropbox.com/s/z3tedek2i7ejen8/7%20%20-%20CAC%20PAYBACK.png?dl=0) to become profitable. SaaS world changes that a bit - payback periods have accelerated in the - 5X increase in payback under 6 months (see report at #2 below for more on that) CAC payback still really highlights the deep dependency on access to capital to fund a Tech company's growth through these [SaaS Cash Flow Troughs](https://www.forentrepreneurs.com/saas-economics-1/).

Link: https://www.dropbox.com/s/z3tedek2i7ejen8/7%20%20-%20CAC%20PAYBACK.png?dl=0

## 2. GTM

A genuinely meaty report from Mercury Fund - their [2026 Early-Stage GTM survey](https://www.dropbox.com/scl/fi/9qpif13k855nmapmlv6a3/6a1196b5ee0c1a831230fe19_Mercury-Fund-GTM-Report_Final.pdf?rlkey=kmzwillmpma6s26tljw9y2uoy&dl=0) (125+ B2B leaders) covers funnel, unit economics (see CAC above), comp, org design, and AI.Win rates up from 29% to 33%, mid-funnel sliding (Lead-to-MQL 41% to 32%), 4:1 LTV:CAC from 0% to 14% of teams.  ChatGPT adoption at 80%, Claude at 60%, Gemini at 50%, though that may have changed in the months since this survey was completed.

Link: https://www.dropbox.com/scl/fi/9qpif13k855nmapmlv6a3/6a1196b5ee0c1a831230fe19_Mercury-Fund-GTM-Report_Final.pdf?rlkey=kmzwillmpma6s26tljw9y2uoy&dl=0

## 3. AEO

ChatGPT referral traffic to monitored sites[ jumped over 60% in May](https://www.growthunhinged.com/p/is-ai-talking-about-your-product-in-the-right-way), and homepages went from 3.5% of those referrals to \~24%. The key takeaway: your homepage is now read by the LLM and clicked by the human, so it needs to serve both of 'em. Seven tactics in there from Clay, Rippling, Lovable, Ahrefs.

Link: https://www.growthunhinged.com/p/is-ai-talking-about-your-product-in-the-right-way

## 4. BOT SPEND

Agentic AI spending is gonna hit $201.9B in 2026, up 141%, and it overtakes chatbot/assistant spending in 2027 for the first time, [according to Gartner's first dedicated AI forecast](https://softwarestrategiesblog.com/2026/02/16/gartner-forecasts-agentic-ai-overtakes-chatbot-spending-2027). Chatbots peak at $264.7B then decline; agentic grows at 119% CAGR to $752.7B by 2029. Talky Bots to Acty Bots.

Link: https://softwarestrategiesblog.com/2026/02/16/gartner-forecasts-agentic-ai-overtakes-chatbot-spending-2027

## 5. HIGH AGENCY

The high-agency test - do you dare to try it? It's the 10x you. [Picture a version of you ten times faster and braver](https://www.thevccorner.com/p/high-agency-success-guide), then do what they'd do today, the scary email, the project you'd close. The gap, apparently, isn't information, it's willingness to take discomfort now.

Link: https://www.thevccorner.com/p/high-agency-success-guide

## 6. PLATFORMS

Check [this report covering platform sales and revenue](https://www.insightrevenue.com/resources/state-of-revenue-2026) - The biggest gap between high and low performers in platform sales isn't how they sell, it's what happens after signature: on whether customers can see value post-go-live, top sellers scored 1.86 vs 3.23 (lower is better) - platform sale fails not because buyers aren't interested, but because sellers answer feature questions instead of the three buyers ask, can it be done, is it worth it, will we survive.

Link: https://www.insightrevenue.com/resources/state-of-revenue-2026

## 7. VENTURE

[Nightmare VC story time](https://www.thevccorner.com/p/vc-horror-stories) - A Sequoia partner passed on Cloudflare early because he didn't think a woman could run a security infrastructure company (now worth \~$80B). A $5M VC commit was cut to $100K on closing day, A GP fell asleep during a Series A pitch, and nobody in the room woke him - way more cringe stuff in there!

Link: https://www.thevccorner.com/p/vc-horror-stories

## 8. PRICING

If you even vaguely sell a wrapper, [don't price AI like you are reselling electricity](https://tomtunguz.com/so-you-want-to-sell-inference/). Price it based on the work it gets done. If a customer brings their own API key and sees raw inference cost on their bill, will your pricing survive? Cost-plus margin breaks, value-based holds, because you're selling outcomes, not tokens. Good example in there - Sierra charges per resolved ticket (and $0 for failures).

Link: https://tomtunguz.com/so-you-want-to-sell-inference/

## 9. NO!

Digging t[his one back from the archives](https://medium.com/free-code-camp/the-most-important-skill-a-programmer-can-learn-9d410c786baf) - every feature you say yes to is maintenance, bugs, and onboarding cost you signed up for forever. So knowing when NOT to code is the skill most of us need to get better at or master.

Link: https://medium.com/free-code-camp/the-most-important-skill-a-programmer-can-learn-9d410c786baf

## 10. CASE STUDY

Ascend COO Omar Ismail [boosted ARR 38% in six months](https://www.bvp.com/atlas/building-a-growth-engine-from-zero-how-one-coo-boosted-arr-38-without-a-growth-team-using-claude-code) without hiring a growth team, using Claude Code to build and run the entire stack. Six months ago: $20M ARR, 95% from word-of-mouth, zero paid acquisition

Taking the week off next week - Happy 4th of July break y'all!

Link: https://www.bvp.com/atlas/building-a-growth-engine-from-zero-how-one-coo-boosted-arr-38-without-a-growth-team-using-claude-code

## POD OF THE WEEK

LOOPS: Didn't you know? [Prompts are out, loops are in](https://www.youtube.com/watch?v=JoXbk2fm7jM) - (a [loop is just an automated prompt](https://x.com/djfarrelly/status/2067677007140278630?utm_source=tldrnewsletter) listening, or on a cron, hook, or goal). In the podcast, they built a self-scheduling PR reviewer and a skills loop that spins up its own validating sub-agents. I copied both of them because they are actually handy.

Link: https://www.youtube.com/watch?v=JoXbk2fm7jM