← Back to case studies
Case study

$1.9M+ active pipeline and 315 interested replies for SpeedSize's AI media CDN

$1.9M+
Pipeline
315
Interested replies
10.15%
Peak reply rate
Category
SaaS
Industry
AI media CDN for premium brands
HQ
United States
Engagement
1 year

About SpeedSize

SpeedSize holds the only patent on an AI media CDN driven by neuroscience. It rebuilds a store's images and videos to load five to ten times faster while preserving one hundred percent of the original quality.

The mechanism auto-fixes Core Web Vitals warnings, cuts file sizes by up to ninety percent, and integrates as a plug-and-play layer on Shopify, Magento, and WooCommerce with no code. The category is enterprise-grade media CDN, head-to-head against Akamai, Cloudflare, and CloudFront. GAP's public POC against Akamai on 200M live requests is the flagship credibility anchor. Allbirds, MÁDARA Cosmetics, Leibish Jewelry, Cozey, Nkuku, and 35+ other premium brands run in the reference library.

The challenge

01
A three-tier ICP, three budgets
1M+, 100K, and 20K-visitor brands each buy differently.
The challenge

Enterprise, mid-market, and entry tiers each support a different deal size and a different register. Mixing them collapses the pitch, so every tier needed its own recipe.

02
A category owned by legacy CDNs
Akamai, Cloudflare, and CloudFront are the default.
The challenge

Reaching a buyer who never considered a specialized media CDN meant naming the exact speed-and-quality gap they were already living with, not pitching a category they didn't know existed.

03
A quality claim doubted by default
"100% quality preserved" reads as marketing copy.
The challenge

The preservation claim gets discounted the first time it is read. Proof had to come from GAP's POC data and the prospect's own measured numbers, not adjectives.

04
An installed base going unused
Prospects installed the app, then never switched it on.
The challenge

A pool of brands had installed SpeedSize's Shopify app but never activated a client ID. A warm, high-intent audience that standard cold outbound treats as cold.

The GTM engineering in works

GTM Play 01
Open with their measured LCP
Every send led with the prospect's real speed score.
10.15% peak positive-reply rateagainst a 1 to 3% category benchmark
The play

We scraped each prospect's actual Largest Contentful Paint score before every send and opened on the number: "your 4.2s LCP exceeds Google's 2.5s threshold."

Personalization stopped at what could be measured, so a healthy-LCP prospect got a different track. This measurable opener carried the book to a 10.15 percent peak positive-reply rate.

Ran book-wide as the default opener
GTM Play 02
Category-calibrated proof
Jewelry heard Leibish; skincare heard MÁDARA.
The play

Every prospect saw the reference closest to their own shelf: Leibish for jewelry, MÁDARA for skincare, Allbirds for apparel, GAP versus Akamai for enterprise.

Named proof from an adjacent brand, never a generic "trusted by hundreds."

GTM Play 03
Test the openers, not just subject lines
Three to four Email 1 variants per campaign.
~2x best opener vs the weakestmeasured by per-variant reply attribution
The play

Each campaign ran a named-executive observable, a role-specific pain frame, an industry-stat opener, and a video-forward hook in parallel, with reply attribution split by variant.

That told us which opener style did the work per tier, not per campaign, so the winner scaled and the rest were cut.

GTM Play 04
The sponsored test as the close
Every campaign closed on a free, no-commitment test.
The play

A free staging AI-CDN, free integration, free PageSpeed audit, and a 14-day money-back test were named as the actual offer, not softeners.

Positioned as risk-free without ever using the word "trial." The buyer walks in knowing they walk out with data either way.

GTM Play 05
Peak-season reactivation
A timing track for warm leads gone quiet.
The play

A separate track built around peak-season timing, launched to past conversations that had gone quiet, acknowledging the earlier thread directly.

Anchored on peak-season pain: 53% bounce after one slow page, with mobile driving over half of peak traffic.

GTM Play 06
Retarget the installed-but-lapsed
App installed, client ID never activated. Warm intent.
~3x the cold positive-reply ratea real intent signal cold outbound ignores
The play

A dedicated track for prospects who installed the app but never grabbed the client ID, opening directly: "saw someone from your team installed our app but didn't grab the client ID."

The setup was already halfway done, so the sequence closed on a free AI-optimization offer and converted well above the cold book.

Results

$1.9M+
Active pipeline in flight, across 99 open deals and 74 unique prospect companies
315
Interested replies generated across three ICP tiers
10.15%
Peak positive reply rate, against a 1 to 3% category benchmark

In their words

Sagi Keen, CEO of SpeedSize
Sagi Keen
"We're getting amazing ICP leads now, great job. Also just got LG.com. That's a mega enterprise. How did you pull this off?"
Free test campaign

Want results like these on your ICP?

Apply below. If we're a fit, we run a real pilot on your ICP and show you the actual replies before you commit to anything.