Executive summary
Meridian sells and installs water treatment, filtration and indoor air systems across nine US states, on a 640-page WordPress site with a 210-SKU WooCommerce catalogue. Marketing owned the pipeline number but not the publish button. A single blog post took 11.5 hours; a landing page took twelve business days; $102,000 a year went to an agency and freelance developers for work the team wanted to do itself. Connecting the company's existing Claude subscription to the site through wptaskify removed the queue without adding headcount.
The challenge
Publishing was a relay race: writer drafts, designer sources images, marketing ops formats, agency uploads, IT approves - five handoffs, 11.5 hours, three weeks of calendar time. 38% of posts shipped with a missing or auto-generated meta description. Technically, 412 open SEO issues (61 broken links, 148 images with no alt text), a 41% Core Web Vitals pass rate, and no audit trail. The backlog grew 61% year over year while headcount stayed flat.
Why wptaskify (vs the alternatives)
Four options were scored against eight weighted criteria by marketing, IT and finance.
| Option | Score / 100 | Verdict |
|---|---|---|
| A · Keep agency | 58 | Preserved the exact bottleneck the project existed to remove. |
| B · Hire a developer | 61 | ~$110k fully loaded; solved implementation, not content production. |
| C · Standalone AI | 64 | Good at drafting, weak at shipping. Output still pasted by hand + a second AI bill. |
| D · wptaskify | 87 | Removed production hours, not just drafting. Governance built in. No second AI bill. Selected. |
wptaskify removed the production hours, not just the drafting hours - publishing, images, meta, schema and internal links happen inside the same instruction, so work lands live instead of in a document. It used the Claude seats Meridian already paid for (no second AI bill, $99/mo connector for up to 20 sites), and governance was built in: approval inbox, automatic pre-edit backup, AES-256 encrypted credentials, and order tools restricted so customer records cannot be deleted.
The four-week rollout
One part-time owner (0.4 FTE), no consultants, no replatform.
| Phase | What happened |
|---|---|
| Week 1 · Discovery | Full crawl + AI SEO Score produced a 412-issue inventory. Approval policy and access map set on staging first, then production. |
| Week 2 · Pilot | Two people, one blog category, five posts end to end. SEO pass across 40 posts. Effort fell to 2.9 hrs/post; rework 31%. |
| Week 3 · Expanded | Landing pages moved in-house (the 12-day bottleneck). Metadata automation across 345 posts. First landing page shipped in 1.5 days. |
| Week 4 · Production | Governance switched on: named approvers, weekly log review. Agency renegotiated $6,500 → $2,275/mo. Go-live on schedule, rework down to 12%. |
What made it hold: the pilot was deliberately narrow - one category, two people, five posts. The team resisted opening access to everyone in week 2, which is the most common reason this kind of rollout stalls.
Results — month 6 vs a 90-day baseline
| Metric | Before | After | Change |
|---|---|---|---|
| Publishing time per post | 11.5 hrs | 1.5 hrs | -87% |
| Landing page turnaround | 12 days | 2 days | -83% |
| Content published | 26 / qtr | 35 / qtr | +34% |
| Organic sessions | 41,200/mo | 52,300/mo | +27% |
| Organic leads | 148/mo | 209/mo | +41% |
| Technical SEO issues | 412 | 47 | -89% |
| Core Web Vitals pass rate | 41% | 92% | +124% |
| Developer hours on marketing | 32 hrs/mo | 9 hrs/mo | -72% |
| Agency spend | $6,500/mo | $2,275/mo | -65% |
| External website operating cost | $8,500/mo | $2,675/mo | -69% |
All five SMART goals were met or exceeded: publishing time -87% (target -70%), developer hours -72% (target -60%), content +34% (target +30%), Core Web Vitals 92% (target >85%), external cost -69% (target -50%).
The ROI model
The model counts only hard cost reductions and recovered labour. Revenue from lead growth is shown separately and excluded, because multi-touch attribution is harder to defend in a finance review.
| Measure | Value |
|---|---|
| Year 1 benefit, ramped | $80,501 |
| Year 1 investment | $18,098 |
| Year 1 net benefit | $62,403 |
| Year 1 ROI | 345% |
| Payback period | 5 months |
| Year 2 net benefit, run rate | $105,192 |
| 3-year NPV at 12% discount | $218,200 |
Even in a conservative scenario (agency reduction 40% instead of 65%, no capacity credit) the project pays back inside the first year (141% ROI, 8-month payback). The variable with the most leverage is the agency line, so that renegotiation should be sequenced early.
Five pitfalls, and what fixed each
A narrow pilot (one category, two people, five posts) is small enough to fix problems privately and real enough to prove the case. Too many untrained users early is the most common reason this kind of rollout stalls.
The month-2 remediation across 412 issues made everything published afterwards perform better. Doing it later would have delayed the traffic curve by a quarter.
The single highest-leverage hour in the project. It cut editorial rework from 31% to 7%.
Three weeks of overlapping cost was avoidable.
Effort per post improves immediately; volume, traffic and leads lag two to three months because they depend on trust. Set that expectation in writing at kickoff.
In their words
"The change I did not expect was psychological. When you know you can publish something today, you start thinking about different things. We stopped planning around the queue."Marketing Director · Meridian Home Systems (composite)
"I was the most sceptical person in the room. What convinced me was the backup running before every file edit and the approval inbox holding anything that touches commerce."IT Director · Meridian Home Systems (composite)
"We spent $102,000 a year to publish two posts a week. We now publish three, ship landing pages in two days, and the capability sits inside the company now."Chief Executive Officer · Meridian Home Systems (composite)
Try the same workflow on your site
Connect your WordPress, WooCommerce or Shopify site and your own Claude or ChatGPT in about two minutes. Free forever plan, no credit card.