Design agency: rebuilding outbound after losing a 30% revenue client

The situation

The client was a 22-person product design agency focused on B2B SaaS. Their top client — a mid-market SaaS company they’d worked with for nearly 4 years — acquired an in-house design team and ended the retainer with 60 days’ notice.

That client was 30% of monthly recurring revenue. Loss was ~$68k/month.

The agency had 11 active retainer clients going in, most between $6k and $14k/month. A handful of project-based clients at lower volume. No outbound motion at all — 100% referrals and inbound from the founder’s personal brand (a Twitter/LinkedIn presence with ~45k combined followers, mostly product people).

The founder had 60 days to figure out whether to:

Make layoffs to match the reduced revenue, or
Try to backfill the revenue through new client acquisition

He had some runway. He didn’t want to lay people off if he could avoid it. The question was whether outbound could meaningfully contribute inside 90 days — the payroll cushion after which layoffs would become inevitable.

I told him upfront: 90 days is aggressive for outbound to replace $68k/month in retainers. A normal outbound rebuild at this scale takes 5–7 months to pay back. But if his personal brand and existing customer base were as strong as they looked, we might compress it. No promises.

The bet we made

Two things shaped the approach:

Use his brand as the atomic unit of outbound. Not agency-sent. Founder-sent, referencing his own publicly visible work. Peer-evidence opener variant (see Signal-based opener library).
Go narrow on ICP to maximize close rate. His brand resonated most with SaaS companies at 50–200 employees who had hit product-market fit and were trying to level up the design function. Broader ICPs (early-stage, enterprise) converted worse historically. For a 90-day rebuild, we needed the highest-converting segment, not the biggest.

We skipped the usual “widen the funnel” instinct. The goal wasn’t 100 meetings. It was 5 retainers.

Month 1 — Setup and list

Infrastructure

Minimal by design:

One sending domain (a lookalike of the agency primary), warmed over 3 weeks.
DMARC at p=reject from day one — we had no legacy senders on the lookalike, safe to do.
Instantly for sending, nothing fancy.

List

650 target companies. Built manually over 8 days. Criteria:

US/UK/Canada-based B2B SaaS companies, 50–200 employees
Visible sign they were investing in design: recent product redesign (check press + public changelog), recent design leadership hire (LinkedIn), or recent public conversation about product design on a podcast or blog post by a company leader
Founder or CEO actively on Twitter or LinkedIn (verified we could reach them personally)

The “founder/CEO on social media” filter was unusual but deliberate. The founder’s entire outbound approach was peer-to-peer, and that only works if the recipient is someone who exists in the same conversation. Targeting a CEO who had zero social footprint meant the “saw your recent post about X” opener had nothing to work with.

Enrichment: Clay waterfall to get verified email for the founder/CEO. Fill rate was ~78% on this narrow list (higher than a broader list because these people are more publicly findable).

Final usable list: 510 rows.

Month 2 — Sending and the surprising pattern

The opener was simple and worked the first time:

[First name] — saw your [specific post/talk/podcast about product design].
The point you made about [specific observation from it] is something
I've been working on a lot with [peer type: founder of a SaaS in your
segment]. Won't pitch you — but I wrote a short take on the same question
here, in case relevant: [link to founder's personal essay on the topic]

If it sparks anything, always open to a conversation.
— [founder first name]
[agency name and one-line description]

Key features:

First sentence is a real, specific reference to the recipient’s own public work.
No pitch.
The “CTA” is a link to a piece of the founder’s own writing on the same topic — not a booking link, not a case study, not a demo. Essay.
Sign-off is the founder’s first name, with the agency context as an afterthought.

Reply rate landed at 6.1% positive across the first 400 sends. Higher than I expected. Two reasons it worked:

Prospects read the essay. The value-first CTA meant we got a reply only from people who had actually engaged. When they replied, they were already half-qualified.
Founder credibility stacked. The email was from the founder. The essay linked was by the founder. The follow-up conversation, if it happened, would be with the founder. No hand-off, no SDR, no bait-and-switch. Prospects noticed.

The unexpected pattern: ~40% of positive replies were from founders who explicitly said “not shopping for a design agency right now, but your essay landed — can we just talk?” Those non-commercial conversations converted to retainer engagements 6 months later at surprisingly high rates. Some we could attribute directly. The funnel has a long tail.

For the 90-day rebuild window, though, we cared about the ~60% of replies that were commercially-minded. Those became meetings.

Month 3 — First retainers

By end of month 3:

1,100 sends completed across all three months.
63 positive replies.
38 meetings booked.
5 retainers signed.

Retainer sizes:

$14k/month (US SaaS, Series B, full product design retainer)
$11k/month (UK SaaS, Series A, product design + brand sprint)
$9k/month (US SaaS, recently funded, design leadership + IC support)
$7.5k/month (US SaaS, bootstrapped, product design on a monthly cadence)
$6k/month (Canadian SaaS, recently funded, design audit + ongoing)

New MRR added: $47.5k/month. 70% of the lost client’s revenue, inside 90 days.

Not full replacement, but enough that layoffs were off the table — the remaining gap was coverable from cash reserves for 2–3 months while the program continued to produce.

Month 4 — Running it forward

Month 4 added two more retainers:

$13k/month (larger US SaaS, referred through one of the month-3 conversations)
$8k/month (US SaaS, outbound-sourced, moved fast)

MRR added across 4 months: $68.5k/month, matching the lost client almost exactly.

Active retainer count: 11 → 16.

Two of the 16 were at larger contract values than the lost client had been, so the agency was actually in a slightly stronger revenue position than before the loss.

What was specific to this engagement

Several things that won’t replicate everywhere:

The founder’s personal brand was real. Not claimed — actually read and engaged with in his ICP. If he’d had a smaller or less-relevant audience, the peer-evidence opener wouldn’t have carried as well.
The founder could write one substantive essay per month. The emails linked to actual writing, not a landing page. Founders who can’t write (or hate writing) can’t run this motion.
The ICP was narrow but TAM was large enough. 510 rows against a $68k/month revenue target was tight. Had the ICP been narrower, we’d have run out of list.
The retainer economics were healthy. Agencies running on low-ACV, high-volume retainers couldn’t recover this fast from a single-client loss — they’d need to sign 15–20 new retainers, not 7.

What was generalizable

For any agency dealing with unexpected revenue loss and considering outbound:

Founder-sent is 3–5x more effective than agency-sent when the founder has credibility. Especially in service businesses. Don’t dilute the strongest sender.
Narrow ICP beats broad ICP in a time-constrained rebuild. Higher close rates matter more than volume when the goal is 5 retainers in 90 days, not 100 meetings.
Value-first CTAs work if you have genuine value to offer. A link to real writing by the founder is value. A link to a blog post regurgitating industry stats is not.
Brand-side revenue loss is a forcing function, not a catastrophe. Every agency I’ve worked with who lost their top client and used it as a forcing function to rebuild outbound came out of it stronger.

The honest caveats

Things that didn’t go smoothly:

Month 1 produced zero retainers. The pipeline was building but no meetings converted in the first 30 days. The founder had 4 weeks of visible pressure to stop the program. He didn’t, which is the only reason any of this landed.
The first 2 retainers signed at lower values than we’d quoted — the founder felt market pressure and discounted to secure the relationships. That’s a common stress response to revenue loss; it’s also a recoverable mistake.
One retainer from month 3 churned at month 6. Was a poor fit that we’d sensed but accepted. Net: the 4-month number overstates 12-month persistence slightly.
The essays the founder linked to in emails took ~6 hours each to write. We counted this as essential program cost, not as “free content marketing.” Founders who don’t have time to write essays would need a different CTA.

The broader takeaway for agencies

Losing a top client is survivable. The instinct to panic-hire a BDR and blast a generic list usually makes things worse — it fills the pipeline with wrong-fit prospects, dilutes whatever brand equity the agency has, and burns domain reputation in a panic send cycle.

The opposite instinct — go narrow, go founder-driven, go slow with messaging — produces better results faster for time-constrained rebuilds, because the narrower motion converts higher even if it generates fewer raw meetings.

90 days is aggressive. Not impossible. Requires the founder’s personal leverage to be the primary engine, and requires discipline on ICP selection.

The client in this case study had both. That’s why it worked. If the founder had been a behind-the-scenes operator with no public credibility, the playbook would have been different — probably a 180-day plan with SDR-sent outbound and case-study-led messaging.

Every rebuild is specific to the agency’s existing assets. The template: find the strongest lever you already have, concentrate outbound around that lever, cut everything that dilutes it. For this agency the lever was the founder’s brand. For yours it might be a specific case study, a specific vertical, a specific service specialty. But it’s there — every agency has one. Rebuild around it.