Search Engine Optimization

There was a stretch when the only way to get a company noticed was to rent the attention. You bought a page in a newspaper or a banner on somebody else's site, and your name was visible for exactly as long as the invoice ran. That arrangement works while buyers have nowhere to look except the people selling to them.

The search box broke that assumption. Somebody choosing a tool for their team now types something like "automate meeting notes," reads the dozen or so results on the first screen, and builds a shortlist out of them. Nine out of ten searchers never reach the second screen. A page sitting at position eleven is, for practical purposes, a page that does not exist.

Search traffic also behaves differently from traffic that arrives any other way, because the person typed the question themselves. Someone who clicked a banner has not yet decided what they need. Someone who searched "automate meeting notes" has already compressed their problem into one sentence and is hunting for the answer. That is why the same hundred visitors can convert at wildly different rates depending on where they came from.

The complication is that a program decides the order, not a person. A search engine runs on three pieces: a crawler that walks the web and pulls pages down, an index where those pages are stored, and a ranking system that picks from the index and puts results in order when a query arrives. A page that looks well made to a human and a page those three can read are not the same page.

At Pagedeck the gap showed up as a single image. Ravi published the product overview exactly as the design file came out of the tool — one flat picture with all the copy baked into it. Six months later the page ranked for nothing except the company name. Human eyes saw a screen full of words. The crawler carried away a filename. Nothing was wrong with the copy itself; it simply was not text as far as the machine reading the page was concerned.

Search engine optimization (SEO) is the work of adjusting a site and its writing to fit those rules. What separates it from advertising is where the cost sits. Ads bring people in while money goes in and stop the day it stops. A page that has earned its position keeps bringing people in during a month when the budget is zero. The bill arrives as time instead. A new article usually takes three or four months to climb anywhere near the top, and a whole site takes six months to a year to start producing steady traffic. Making something worth finding is the job of content marketing, which produces writing people actually want to read; SEO is what lifts that writing up the results page.

The work underneath splits four ways: the areas you actually touch, the screen where the results appear, the published standards a search engine uses to sort pages, and the settings where search happens somewhere other than the open web. The split matters because rankings stall for very different reasons. If you picked the wrong words to target, no amount of page editing helps. If the page never gets crawled, rewriting it is wasted effort.

Splitting the work this way also settles who owns what. On a small team one marketer covers all of it, but the four branches pull on different skills, and the moment a second person joins the line usually falls between the writing side and the site side.

Areas are where hands-on work happens, and there are four of them.

1. Keyword research — finding the phrases people really type and deciding which ones to chase. What you call a "collaborative document solution" your customer searches for as a "meeting notes app." A phrase with 30,000 monthly searches is usually too crowded for a young site to break into, while a phrase searched 200 times a month by someone who already understands their problem brings fewer visitors and far more signups. Choosing the mix is the decision this stage makes.
2. On-page SEO — adjusting what sits inside the page to match the phrases you chose. The title line shown in results, the description under it, the heading structure of the body, the alternative text on images. Change a title from "Feature overview" to "How to write meeting notes automatically" and the set of queries the page can answer changes with it.
3. Technical SEO — clearing away whatever stops a search engine from reaching and reading the site. Slow pages, settings that block the crawler, and the same content living at several different addresses are the usual culprits. This is the branch where rankings improve without a single word being rewritten.
4. Off-page SEO — building reputation outside your own site. When another site cites your page and links to it, a search engine reads that link as a vote. Unlike the other three, you cannot simply do it; somebody else has to decide your page was worth pointing at.

The results side is where all of this becomes visible. The search engine results page (SERP) is the whole screen a query returns. Reading the rank number alone misleads you, because the layout of that screen changes from query to query. Put four ads at the top, a map under them, and a box of related questions under that, and the first organic link sits well below the fold. That is how a page reaches position one and brings in no more visitors than before.

Evaluation standards describe what a search engine says it is looking for. E-E-A-T stands for experience, expertise, authoritativeness, and trust — four qualities Google publishes as guidance. It is not a score that feeds directly into ranking; it is the instruction given to the human raters who judge quality. Still, it explains which of two articles on the same subject ends up higher. An unsigned post and a post by someone who has done the work for three years and puts their name on it separate here.

Settings are branches where the stage itself differs, so the rules differ too.

1. Local SEO — queries with a place attached, like "accountant near me." Rankings turn less on what the page says and more on the business listing on the map, the posted hours, and the number of reviews.
2. App store optimization (ASO) — the stage is an app marketplace rather than the web. There are no links to trade, so off-page work disappears entirely, and the icon and first screenshot decide installs instead.

The four feed each other in order. Keyword research settles what to aim at, page-level work builds something that answers it, technical work opens the road so it can be read, and outside reputation pushes the page up in competitive queries. You then check how the position actually looks on the results screen and use the published standards to decide what to fix next. Leave one branch empty and the rest spin. A team that writes beautifully but blocks the crawler ships a year of articles that never appear anywhere. A team that fixes every technical issue without choosing target phrases ends up ranking first for sentences nobody searches.

"We're at position three for that query and the click-through rate is two percent. Let's rewrite the title and description before we touch the article."

The ranking is fine but nobody is clicking, so fix the two lines that show on the results screen first.

"That page isn't in the index yet. Rankings are a later conversation."

The search engine has not even added the page to its list, so editing the content changes nothing.

"This is a three-month play, so it won't show up in this quarter's numbers."

The payoff lands outside the quarter and should not be measured against it.

At Pagedeck this conversation runs long once a quarter. Ravi lists search traffic by query, and two or three phrases out of a dozen usually account for half of everything. Mia reads the same list from the product side. If the biggest source of traffic is "meeting notes template" and visitors from that article sign up at one percent, the article is ranking well for a phrase typed by people who are not shopping for a tool yet.

Theo joins when the team decides what to fix, because some of the low rankings have nothing to do with writing. Last quarter four hundred help-center pages were never being crawled at all, and the cause was a setting that required a login to open them. Six weeks after that setting came off, the help center pulled in more visitors than the product overview. Nobody wrote a new article.

What usually splits the room is sequence. There are always dozens of things to fix and three people looking at different ones. Pagedeck puts search volume and current position in one table and starts with the phrases sitting between positions eleven and twenty. Those are already at the edge of the first screen, so a small push produces traffic, and the effort is nothing like dragging something from position thirty to the top. Anything below that line goes on a separate list the team revisits once the near-misses run out.

The most common misreading is that this work means stuffing keywords into articles. The visible activity is word-level editing, so the assumption comes easily. In practice, repeating a phrase twenty times on one page stopped working around 2010, and today it pushes a page down. What a search engine measures is whether the person who asked the question got an answer and stopped looking. Left uncorrected, this belief produces a pile of unreadable articles: rankings never move, and the few visitors who do arrive leave after three lines. The cost compounds, because those same pages are the ones a search engine samples when it decides how much to trust the site as a whole.

The other frequent mistake is treating it as a one-time project. Fixing a site genuinely is the kind of job that finishes, so it looks that way. But a ranking is not a score attached to your page — it is a position among every page chasing the same phrase. Stand still and a new competitor moving up pushes you down. The ranking rules themselves change several times a year. A team that leaves a two-year-old article alone usually discovers the slide only after traffic has halved, and winning the position back takes longer than earning it did. Planning for that is mostly a matter of reserving capacity: budget part of every quarter for revisiting pages that already rank instead of spending all of it on new ones.