The Support
Field Manual
Everything from the staff meeting, distilled into what you actually need at the desk: how to answer the hard tickets, what never to do, and what's coming next.
SEO is a 3–6 month game
Almost every "no results" complaint is an expectation problem. Set the timeline early and often.
Diagnose the real issue
Don't defend the tool — check their backlink & competitor gap on the Backlinks page. The gap is usually the answer.
Never offer refunds first
Only act if the customer explicitly asks. The policy lives in the T&C. Money is the business's oxygen.
Never touch their website
No WordPress, no DNS, no logins, no clicking their links. Security risk + they'll blame us for everything after.
Links are the moat
Whatever happens to SEO, link building stays valuable. It's the metric that can't be faked. Upsell it with confidence.
Use long, AI-written answers
For confused customers, a thorough explanation does the work. Length + substance ends the back-and-forth.
- Lead with the timeline: SEO takes 3–6 months, not weeks. It's not the tool — it's the clock.
- Open the Backlinks page and compare them to their competitors. Show the gap: they may sit at DA 10–20 while competitors are at DA 30–40.
- Explain that DA is a logarithmic scale — DA 20 is roughly 10× stronger than DA 10. They're often "years behind."
- Frame AutoSEO as the fastest safe way to catch up: going faster would trip Google's spam filters.
- First, check if it's even real. Small swings are noise — 300→250, or 3→2 visitors means nothing. Only a 20–30% drop sustained over 2–3 months is a real signal.
- Explain the Google "testing dip": when a site suddenly improves (more articles, more links, longer content), Google may temporarily lower traffic to test if it's genuine or gaming the system.
- If the owner keeps going, Google concludes it's legit and traffic grows. If they panic and stop, that confirms gaming and it stays down.
- The worst outcome is a customer quitting over a 10% dip. Keep them publishing. (Back it up with the Google "rank transition" patent — full explanation in Product Knowledge below.)
- Never offer a refund proactively. Only respond if they explicitly ask. The 7- & 30-day guarantee already lives in our T&C.
- We're genuinely fair (top-percentile) on refunds — but don't volunteer it. Many customers won't even bother asking over $149.
- Someone who used 2–3 months of service (especially backlinks) and wants it all back is hacking the system, not making a fair claim.
- New policy (being rolled out): a refund requires deleting all articles + we remove all links. Not yet enforced — flag it if customers find it confusing.
- This is by design, not a bug — it's called topical authority.
- About 1/3 are "pillar" articles (long, deep), and the rest are supporting articles around them. You need 30–40 articles on one topic to signal Google you're an expert.
- Reassure: users don't read every article, it's not duplicate content, and it helps rankings rather than hurting them.
- Caveat: for a very narrow niche, genuine repetition can creep in — note those cases, but the strategy itself is intentional.
- Hard no. Never accept or request access to a customer's website, DNS, or accounts.
- Two reasons: it's a security risk, and once we touch their site they'll blame us for anything that breaks for the next 6 months ("I uploaded the plugin and my site went down → it's your fault").
- Also don't click links or downloadables customers send to "get into" their site (a real case we've had). No upside, all downside.
- Instead: guide them, or point them to their web developer. If they have a site, someone manages it.
- We have three explicit consent points: (1) choosing annual vs monthly on our page, (2) entering the card on Stripe, (3) the day-3 trial email that states it auto-renews.
- Reality: people still miss it. That's exactly what the 7- & 30-day guarantee is for — if they feel misled and ask, process it.
- Context: an A/B test now lets users choose annual or monthly, so some genuinely selected annual. (This test lifted yearly sign-ups from 3% → 10%.)
- We try to save and educate every customer — but we don't have to save everyone. Some are simply too difficult or too low on background knowledge.
- Give the honest, truthful explanation once (a long AI-written reply if it helps), then don't burn more time on it.
- Don't take it personally. A genuinely impossible customer is roughly 1 in 100 per week — not a reflection of your work.
✓ Always do
- Diagnose the real issue — links & competitor gap, not "the tool is bad."
- Set the 3–6 month expectation early in every conversation.
- Upsell with confidence when you're 100% sure it helps (backlinks, more credits).
- Write thorough, AI-assisted explanations for confused customers.
- Open a ticket for anomalies; route backend questions to Pete.
- Verify the backlink price before quoting it to anyone.
✕ Never do
- Don't offer a refund unless the customer explicitly asks.
- Don't take or request access to any customer's site, WordPress, or DNS.
- Don't click customer-sent links/files to "get into" their site.
- Don't quote the backlink price from memory — it may be €299, not $249.
- Don't let customers quit over an insignificant traffic dip.
- Don't argue with the 1-in-100 impossible customer — answer once, move on.
💡 The value story
Upsell ladder → DR 20 / DA 40 links → more backlinks → more credits. Push the backlink package especially to canceled or unsubscribed users.
Link building is the moat
SEO is changing fast — AI recommendation engines mean pure article generation may not be enough long-term (the product may pivot). But link building stays valuable for years because it's the hardest metric to fake.
Moz DA and Ahrefs DR can be gamed with cheap spam. The real domain rating that Google & Bing measure cannot. That's why links endure — alongside genuinely good content.
Moz DA vs. Ahrefs DR
We use Moz DA (more reliable). Moz only updates every 60–90 days — crawling the whole web is enormous. Small sites (<10 DA) can take 2–3 months to reflect new links; we re-check every 1–2 weeks.
We won't switch to Ahrefs — it would scramble every user's numbers. Example 358 links built in 2 months, Moz had discovered just 1 → DA barely moved.
The Google "rank transition" patent
When a site changes a lot, fast (many new articles, rapid backlinks, longer content), Google can delay or briefly reverse the ranking gain for weeks — the patent says up to ~70 days. It's a deliberate "rank transition" built to catch manipulators, who tend to panic and change tactics. Genuine owners who keep publishing ride it out and recover — so quitting is the real mistake.
Send the customer the source:
US 8,244,722 B1 — “Ranking documents”
US 8,924,380 B1 — “Rank transition function”
US 7,346,839 B2 — “Information retrieval based on historical data”
Plain-English explainer (SEO by the Sea)
⚠ Don't over-claim: the patent is an anti-spam mechanism, not a literal "test for beginners." Frame it as: "Google watches sudden changes and there can be a temporary swing for a few weeks — steady, genuine work is what wins."
Topical authority
~1/3 pillar articles + supporting articles, 30–40 per topic, is how you prove expertise to Google. Repetition here is the strategy, not a flaw.
The 3–6 month clock
A qualification question asks how fast they want results; if they pick "right away," a follow-up confirms they understand SEO takes 3–6 months. Month-1 and month-2 emails will reinforce it before each renewal.
Small & medium business owners — not SEO pros
People who don't have the time, knowledge, or desire to do SEO themselves. It's a far bigger market than professionals who already know everything.
The "clueless" customer is an opportunity, not a nuisance
30–40% of sign-ups had never done anything in SEO before. Converting non-users into users is the huge prize — so be patient with beginners.
Answers live where the question is asked
Info sits right on the relevant page (e.g. the backlink page is packed with it) rather than in a separate help hub — for now.
Feedback is never about blame
From the founder: feedback targets making things better, never calling out who did something wrong. No one's hunting for mistakes.
AI support assistant Coming to you first
An internal chatbot with full backend access — Stripe, databases, questionnaires, all docs. ~2–5 min per query. You'll be able to ask things like "what links did Moz recently discover for this site?" Rolling out to support, then users (inspired by RankMail's integrated "Ask AI").
Disputes are now manual Live now
Automatic Stripe dispute-refunding was disabled ~1 week ago — it was auto-refunding ~90% of disputes invisibly and enabling abuse. Keep disputes under 1–2% or Stripe can ban the account. 15 open disputes are being fought with AI-generated evidence (usage logs, IP-vs-card match, article screenshots).
Lifecycle renewal emails In progress
New month-1 and month-2 pre-renewal emails will hammer home the 3–6 month message and the competitor comparison, to reduce early churn.
Marketing reset → expect more volume Heads up
The US PPC agency and UK CRO agency (April) both failed and were fired, causing a new-customer dip — now recovering. Testing Facebook, starting Google; Hungarian cold email is working. Expect 2–3× more users soon — and a heavier support load.
The trial/exit email claims "co-founder Mihály is here right now… use the chat now" — a leftover from an old experiment. It's no longer true and misleads customers.
The white-label page still renders a normal AutoSEO page — branding is visible all over it, so it isn't actually white-labeled. One paying tester is affected.
When the customer activated the Enterprise plan, her other active subscriptions were deactivated.
The ~$249 backlink-only package was visible only to subscribers — but it should reach canceled / non-subscribers, and be made prominent on the dashboard.
Support can't tell customers when the competitor list / DA last refreshed, or which metric is shown (we use Moz DA, but the UI isn't clear). Recurring complaint: "hasn't updated in weeks."
New policy: a refund requires deleting all articles + we remove all links — but the system to do this doesn't exist yet. Founder is still gathering feedback on whether to keep the policy.
Unclear whether rewrites are allowed after cancellation (trial = 5/article, active = unlimited, after cancel = ?). Routed to Pete.