// case study 01 · Talent assessments · US / Anglo-Saxon
0 → 1.2M organic users a year. On $7k of backlinks.
We built Gyfted's organic engine from scratch — programmatic SEO at scale, fused with free, ungated psychometric tests that satisfied search intent. Here's the whole curve: the explosion, the peak, and the decline we'll be just as honest about.
1.2M
Users / yr (GA peak)
~42.6k
/mo organic (Ahrefs peak)
$7k
Total backlink spend
#1.7
Avg pos · "spotify personality test"
// the curve — monthly organic traffic
// the brief
A psychometric-assessment platform with open-access tests. The job: own the enormous long tail of "personality test / strengths / character" intent in the US — at a scale no editorial team could hand-write.
// the play
- [→]Programmatic at scale — structured data → LLM rewrite → indexed fast.
- [→]Product-led — free, ungated tests that ARE the search answer.
- [→]90/10 — 90% research & targeting, 10% generation. ~$7k links total.
// proof — live ranking wins (google search console, trailing 12mo)
| Query | Position | CTR | Clicks |
|---|---|---|---|
| spotify personality test | 1.7 | 53.6% | 2,817 |
| strengthsfinder test free | 4.2 | 11.4% | 2,880 |
| dexter test | 5.7 | 3.6% | 2,340 |
| strengthsfinder test | 4.4 | 7.5% | 2,286 |
// The honest part — what the decline taught us
From early 2024 the curve falls — hard. Two causes: Google's Helpful Content updates re-weighted thin programmatic pages, and once the team's focus moved on, the content stopped being maintained. Programmatic SEO is not "build once." It decays without freshness, pruning, and a product moat underneath it.
What we'd defend with now: Ruthless pruning of dead pages, freshness loops, real E-E-A-T signals, and tying every programmatic page to live product utility so it earns its ranking. We measure SEO like psychometricians — the decline is data, not an embarrassment.
// what the data actually says
53.6%
CTR at position 1.7 on "spotify personality test." The page IS the answer — product-as-content owns the SERP, not keyword stuffing.
$7k
drove a ~42.6k/mo peak. Links were never the lever — supply of indexable, intent-matched pages was. Most startups overpay for backlinks 6×.
HCU
The fall tracks Helpful-Content-update timing, not seasonality. Algorithmic and predictable — which means it is defensible.
// takeaways you can steal
- 01 Spend 90% on research & targeting, 10% on generation. AI makes pages cheap; knowing which pages to make is the whole game.
- 02 Free & ungated beats gated when the product itself is the search answer. Don't put a form in front of the intent.
- 03 Programmatic without a product moat decays. Pages that only exist to rank get pruned by HCU — tie every page to live utility.
// is this play right for you?
works if
- You have templatable supply — data, profiles, locations, SKUs.
- Your product can fulfil the search intent, not just describe it.
- You can wait for compounding (quarters, not weeks).
skip it if
- You have no proprietary or structured data to template from.
- Your intent is purely one-off / transactional.
- You need overnight results — this is a compounding game.