How a Perth Picture Framer Turned 500 Dormant Customers Into 361 Conversations and a Full Inbox of Hot Leads
A reactivation pilot for a Perth Picture Framer Capaign June 2026.
Results Summary
MESSAGE SENT
500
CONVERSATIONS
361
REPLY RATE
72%
BUYING SIGNALS
79
GOOGLE REVIEWS
26
BOOKING
3
COMPLAINTS
0
(1 spam, auto-handled)
Testimonial
"Wade's SMS campaign reactivated dozens of past customers I hadn't heard from in months. Our Google reviews are up, the phone's ringing more, and the whole thing runs itself in the background. Honest work - recommend."
- Glen, Owner,
🧔 CLIENT SNAPSHOT
Industry: Custom picture framing (bricks-and-mortar retail + services)
Location: Perth, Western Australia
Database Size: ~5,000 lifetime customers, pilot drawn from top 300 multi-purchase recent contacts
Campaign Period: About two weeks (17 to 30 June)
Service: ReviewProsper AI Database Reactivation (SMS)
Operator: Senior framing shop owner with 25+ years of trade.
⚠ THE PROBLEM
A specialist Perth picture framer had built ~5,000 lifetime customers over twenty-five years of trading. Beautiful work. Loyal base. But the database was functionally dormant, sitting in a POS system with no systematic way to bring people back.
The problems were familiar:
No email list to speak of: past customers had given names and phones at the counter, not opt-in emails
Walk-in traffic was the main demand driver: when foot traffic was slow, revenue was slow
Repeat purchases happened by chance: when a customer happened to have something else to frame
The owner was the bottleneck: every customer relationship lived in his head, no systematic follow-up
Marketing tools didn't fit: email platforms felt too distant, agencies wanted retainers without proof
The hidden cost: customers like "S.B." (Canvas Dot print framed in January 2025) had cricket shirts at home she wanted framed. "P.K. & S.K." (Mermaid print framed in December 2024) had their next piece ready to bring in. "F.S." (Larry Mitchell canvas in March 2025) had two more canvases lined up. None of them had been contacted since their last purchase. All of them said yes when we reached out.
⚙ THE SOLUTION
ReviewProsper deployed a three-tier AI SMS reactivation campaign built on the principles proven in our health-and-wellness pilot, now adapted for a services business with appointment-based selling instead of e-commerce checkouts.
1. Hyper-Personalisation Anchored in the Past Purchase
Every opening message named the exact past framing job and last purchase date:
Recognition triggered instantly. People replied "Yes that's me" because they remembered the artwork, not just the brand. Recall hit rate on first send: 66%.
2. A Trust-First "Two-Text" Opener
Sixty seconds after the introduction, every contact received a second message a brand-safety promise designed to convert hesitation into trust:
Result: customers who didn't want the contact used the word we taught them and exited cleanly. Zero ACMA hits. Zero complaint escalations. 3.3% opt-out rate, entirely through the channel we designed for them.
3. AI-Driven Conversations With a Hard Escalation Rule
"Emma", the AI persona held natural, services-appropriate conversations:
- Confirmed identity warmly
- Surfaced the right offer per customer track (free AR Glass upgrade for past glass customers, 15% off for past canvas customers)
- Answered straightforward questions inline[3:15 PM]- Never guessed. When a customer asked anything outside her playbook a specific artwork question, custom pricing, special handling, Emma routed straight to the owner with the locked phrase: "Good question, let me get the details from Glen and have him call you back."
- Asked for Google reviews from customers who praised the work
- Booked consults directly via a calendar link
Critically: the AI knew its limits. 5 conversations were escalated to the owner for personal handling every one of them a real human moment, not a missed opportunity.
4. Tiered sends with learning between
The campaign ran across three tiers, allowing each batch to refine the next:
| Tier | Date | Sent | Replied | Reply Rate |
|---|---|---|---|---|
| Tier 1 | Tue 17 June 2026 | 100 | 71 | 70% |
| Tier 2 | Mon 22 June 2026 | 100 | 80 | 78% |
| Tier 3 | Tue 23 June 2026 | 100 | 78 | 77% |
| Tier 4 | Tue 30 June 2026 | 200 | 132 | 66% (still climbing at time of publish) |
Between tiers, the offer-reveal copy was refined from premium positioning ("museum-grade, 99% UV protection") to descriptive feature language ("Anti-Reflective Glass, virtually invisible, brightens colours, 78% UV protection") and interested-buying intent rose markedly in Tier 3.
🛎 THE RESULTS
⏲ What Glen Saw the Day After Tier 3
Within hours of the final send, the owner reported back:
he'd had to cut a coaching call short because his phone wouldn't stop ringing. Customers from the SMS were walking in. Booking the calendar themselves. Calling to ask about specific pieces. The shop was busier than it had been in months.
That's the operator-level signal no single metric captures: the campaign produced more demand than the owner could handle in real time.
| Metric | Result |
|---|---|
| Total contacts messaged | 500 |
| Total conversations created | 361 (72% reply rate) |
| Explicit buying signals | 79 customers |
| Direct bookings via AI link | 3 confirmed consults |
| Google reviews directly attributable to the campaign by name and timing. |
26 in about two weeks (17 to 30 June). Verified by cross-matching reviewer names to the SMS pilot list. His profile sits at 108, publicly verifiable. |
| Escalations to the owner | 5 each a real callback request |
| Opt-outs | 10 3.3% — all clean, via the “reply delete” promise |
| ACMA / regulatory hits | 0 |
| Negative-tone replies | <2% of total |
⭐ The Pipeline Snapshot About Two Weeks (17 to 30 June)
Of the 361 customers who engaged:
3 bookings already locked into the calendar without owner intervention
79 interested-buying signals sitting in the inbox, ready to convert with personal follow-up
26 Google reviews we attribute to the campaign by name and timing. His profile sits at 108, publicly verifiable.
Multiple phone consults already triggered by the owner's call sheet
💰 Cost Per Conversation
The full pilot 500 SMS, AI conversation handling, live monitoring, daily call sheets, full oversight infrastructure operated inside the client's existing retainer. No additional ad spend. No paid traffic. No new tools.
The conversion math is simple: a custom framing job averages $300–$1,500. One booked consult covers the cost of the entire campaign multiple times over.
The pipeline as it sits 79 interested signals + ongoing reviews + ongoing repeat business represents conservatively $23,000–$50,000+ in retrievable revenue the database was already holding.
💡 Why this worked (and where most reactivation fails)
1. The "Turing Test" Standard
If a customer couldn't tell they were talking to AI, the AI was working. Every Emma message was written in genuine Australian English, warm and unhurried, with first-name address and zero salesy hooks. The reply "the way you approached this was very VERY nice, I didn't feel advertised lol" came from a customer who'd just declined the offer, and asked Emma to pass on the compliment.
That's the bar. Not a chatbot. A human you'd want to text with.
2. Track-Based Offers, Not Universal Discounts
Past glass customers got the AR Glass upgrade. Past canvas customers got the 15% canvas offer. No universal discount code, no race-to-the-bottom pricing. Each offer matched what the customer had already proven they valued.
3. The Two-Text Opener Was a Force Multiplier
The "reply delete" promise didn't just protect brand trust, it converted hesitation into engagement. Customers who would have ignored a single cold SMS read both messages, understood the social contract, and replied at scale. The 68% pilot-wide reply rate is not normal for SMS reactivation. Industry benchmarks sit at 15–25%.
4. Owner-Aware Architecture
Every part of the system was built around the owner's reality, a busy small-business operator who reads on his phone, doesn't want dashboards, and needs the right answer fast. Daily call sheets shipped to his Slack at 7:30am with full conversation context, real phone numbers, and word-for-word opening scripts. He never had to log into a tool. He just woke up to a single document and worked top-down.
5. The AI Knew When to Step Aside
The escalation rule, "never guess, always route to the owner", protected the brand reputation that 25 years of trading had built. Every uncertain moment became a real human conversation. That's the difference between an AI that adds value and an AI that erodes trust.
🗯 Conversations that prove it worked
Numbers tell you the campaign performed. The conversations tell you why it performed, because customers responded to Emma like a person, not like a marketing system. These are unedited replies pulled directly from the SMS oversight log.
The "I didn't realise it was marketing" moments
- A-C.T. (past glass customer, declined the offer)
This is the line that defines the campaign. A customer who said no and then took the time to compliment the approach. That's emotional resonance, not lead generation.
- N.D. (past glass customer)
A customer who welcomed the message and called it "spam" affectionately. That's the opposite of how SMS reactivation usually lands.
The "I'd forgotten I needed you" moments
- P.K. & S.K. (past glass customers booked themselves in for 8 July directly through the link)
- F.S. (past canvas customer, ready to commit)
- J.C. (past glass customer, re-frame opportunity surfaced by the contact)
Three different customers, all of whom would not have come back this month without Emma's text. The framing job was already mentally queued, the SMS just gave it permission to happen now.
The "Glen's work spoke for itself" moments
- P.J. (past canvas customer)[3:15 PM] "Hi Emma. The text is fine. You guys framed a couple photographs for me at the end of last year… I was very happy with the work you did. Kind regards, Mark" - M.M. (past glass customer - went on to leave a Google review the same day)
You did four of those oak frames for me along with re-stretching the canvases, as I had brought those artworks over with me from England. The frames are really beautiful too! 😁 KarenG - K.G. (past canvas customer)
These are the customers who left the 26 Google reviews we attribute to the campaign by name and timing. His profile sits at 108, publicly verifiable. The praise wasn't extracted, it surfaced naturally because customers genuinely remembered the work and wanted to acknowledge it.
The "Emma knew when to step aside" moments
When customers asked anything outside Emma's locked playbook specific artwork advice, pricing on custom jobs, cricket-shirt framing specials Emma routed to the owner with a single locked phrase:
5 conversations were escalated this way. Each became a real callback. Including S.B., who asked about specials on cricket shirt framing and gave Emma a callback window of "after 3pm." That phone call alone, if it converts, covers the cost of running the entire campaign.
The customer who told us the hard truth
- Customer feedback received in-thread
This is what real conversation looks like. A campaign that only produces happy replies isn't a real campaign, it's filtered theatre. Emma's warmth created enough trust for an unhappy customer to surface a service issue the owner could now resolve. That feedback is more valuable to the business than another five-star review.
The deeply human one
- T.R. (past canvas customer)
An older customer apologising to Emma for being slow to reply, signing off with a kiss. That's not how people talk to bots. That's how people talk to someone they trust.
The compounding effect. Day 8 bump produces reviews 12 days later
The campaign was designed with a one-week silent gap before a single bump message to any contact who hadn't replied. That bump fired this morning, twelve days after the original Tier 2 send, and produced two fresh five-star Google reviews within the hour.
Sarah K. (past glass customer) hadn't replied to her opener on 22 June. The bump fired at 11:10am Perth on 29 June. Her response captures Glen's experience with the campaign perfectly:
She called Glen's shop to verify legitimacy first, then came back to the SMS, told us the wording could be clearer, and left a 5-star review within 30 minutes.
Dave M. (past canvas customer) replied to the same bump shortly after. He confirmed the campaign had the wrong artwork detail in our records, declined the offer, then left a 5-star review and signed off with "You're welcome Emma!!"
These two reviews 2 of today's 13 SMS-attributable total came not from the initial send, but from the patient, low-pressure bump cycle. Most reactivation campaigns abandon non-responders after 24–48 hours. The MPF playbook waits a week, sends one bump, then stops. That patience compounds.
What Glen Said
- Owner, mid-pilot
- Customer K.L. (past canvas customer, on Emma's wrap-up message) [3:15 PM]
- Customer A-C.T. (past glass customer, declined the offer but committed to a future review)
- Customer P.J. (past canvas customer)
Next Steps for the Client
The pilot closes its formal window in early July 2026. Provisional plans include:
1. Phase 2 expansion to the deeper customer database (24–48 month dormant tier)
2. Always-on review nurture flowing from every new in-store interaction back through the same Emma persona
3. Monthly send cadence as part of an ongoing retainer relationship, replacing one-off pilots with sustained pipeline generation
About Review Prosper
ReviewProsper.ai is a senior-led AI database reactivation service. We work with a small portfolio of brands at any one time. We do not sell software. We do not sell self-serve. We build, deploy, and monitor live customer reactivation campaigns end-to-end owned by experienced operators with 15+ years in direct response, paid search, and SEO.
When a campaign works, the operator sleeps better. When the operator sleeps better, the brand grows.
Interested in a pilot for your own customer database? Contact: [email protected]
Case study published 23 June 2026. Final pilot numbers reported as at sync timestamp 23 June 2026, 14:00 AEST. Numbers are live and continue to evolve as conversations mature. Client identity withheld pending public-release approval.
- ReviewProsper · Melbourne, Australia
AI-powered SMS sales system. Built in Melbourne for Australian businesses.
© 2026 ReviewProsper AI. All rights reserved.
AI-powered SMS sales system. Built in Melbourne for Australian businesses.
© 2026 ReviewProsper AI. All rights reserved.