Search Engine Optimization

Start here, because most of what you have been sold is the wrong thing

If you own a business around Cedar Creek Lake, you have been called about this. Someone told you that you are not showing up on Google, that your competitors are, and that a monthly programme will fix it. That call was dialed from a list.

So we will say the unprofitable thing first. For a large share of businesses in Mabank, Gun Barrel City, Kemp, Tool and Payne Springs, search engine optimization is not the right place to put money. The highest-return action for a business in a town of a few thousand people is a correctly configured, verified Google Business Profile with accurate hours and a real number of reviews. That is free. After it, the returns fall off sharply.

The arithmetic is the reason. Search optimization captures existing demand; it does not create it. If the total number of people searching for your trade across the lake towns in a month is a couple of dozen, then capturing every one of them is still a couple of dozen searches.

What follows is what this work is, what Google puts in writing, and where the two diverge from the version being sold.

What Google documents, and what the industry repeats

There is a first-party source for what Google wants, and almost nobody quotes it in a sales meeting. It is called Google Search Essentials, and it used to be the Webmaster Guidelines. It has exactly three parts: technical requirements, spam policies, and key best practices.

The technical requirements are short. Googlebot must be able to reach the page, the page must return a success status code, and the content must be indexable. That is the entry ticket. There is nothing in that document about word counts, keyword density, posting frequency, or any of the other folklore that gets converted into a monthly deliverable.

The second thing worth knowing is that Google publishes a list of named ranking systems, not a list of ranking factors. The named ones include BERT, RankBrain, neural matching, freshness systems, link analysis systems and PageRank. Google lists retired ones on the same page, including the helpful content system, Panda and Penguin. Retired means absorbed into core ranking, not that the underlying concern stopped mattering.

Two habits follow. Anyone still selling recovery from the helpful content update is selling a product Google folded into core ranking in March 2024. And Google is explicit that the page covers named, publicly discussed systems rather than an inventory, so a vendor claiming to know the complete set is claiming something Google has never published.

Crawling, indexing and serving, and where the money actually goes

Google documents three stages, and understanding them tells you which parts of this work are finishable and which are not.

Crawling. Googlebot finds URLs through links, sitemaps and past crawls, then fetches them. A page blocked in robots.txt, or one that returns a server error, does not get in.

Indexing. Google renders the page, works out what it is about, picks a canonical among duplicates and stores it. Indexing is not automatic, and Search Console reports the failure in plain words: Crawled – currently not indexed.

Serving. At query time Google assembles candidates and orders them using many systems at once. There is no dial in that stage anyone outside Google can turn.

The first two stages are largely solvable, comparatively cheap and permanent: a crawlable site, one page per service, correct canonical tags, a sitemap, HTTPS, and pages light enough to load on a phone on a rural connection. That last point is concrete rather than theoretical here, since published broadband data for Mabank puts gigabit availability at 21.9 percent and the fiber footprint at 0.2 percent. The third stage is where budgets disappear, because it is competitive and probabilistic. Google's own account is worth reading before you sign anything: How Search Works.

Why nobody can guarantee a ranking, and what it means when somebody does

Here is the sentence to keep. Google's own documentation states that Google does not guarantee that it will crawl, index or serve your page, even if the page follows Google Search Essentials, and that not every page Google processes will be indexed.

If Google will not promise that a page gets indexed, nobody outside Google can promise that it ranks.

A promise of position one, of page one, or of a fixed number of top-ten placements is one of three things, and it is worth knowing which:

• Meaningless. The contract says one thing in the sales call and something much narrower in the small print, usually with an escape clause about algorithm changes.
• True but worthless. The vendor picks phrases that are winnable precisely because nobody searches them. Ranking first for a phrase with no volume is a screenshot, not a customer.
• Dangerous. The vendor intends to use tactics Google names as spam policies, and that risk sits with your domain, not with the vendor, who moves on.

One question separates them. Ask for the monthly search volume of every keyword named in the proposal, in writing, before you sign. A vendor who works with keyword data has it already. One who will not supply it has answered you.

Measurement is the other half of the problem. Google states that search results are specific to the time, place, device and recent history of the person searching, so checking your own position from your own phone in your own shop tells you very little. Judge this work on calls, on form submissions and on requests for directions. Those are events. Positions are opinions.

The 2024 documentation leak, stated accurately

Almost every pitch since mid-2024 mentions the Google leak, and most describe it wrongly.

Internal API reference documentation for Google's Content Warehouse data store was pushed to a public code repository on 13 March 2024 and noticed publicly on 27 May 2024. Google confirmed on 29 May 2024 that the documents were authentic, while cautioning against assumptions drawn from out-of-context, outdated or incomplete information. The material runs to roughly 2,596 modules and 14,014 attributes.

Now the part vendors skip. The documents contain no source code, no scoring functions and no weights. Nothing in them says how much any attribute counts for, or whether a given attribute is live in ranking rather than deprecated, experimental, or used by a different Google product entirely, and many fields carry deprecation annotations in the leaked material itself. An attribute name existing is not evidence that it is a ranking factor.

What the leak did fairly establish is narrower: attribute names exist that are hard to reconcile with some of Google's past public statements, including click-derived signals, Chrome-derived data, a site-level authority attribute and fields relating to host age. That is a credibility finding about Google's public communications, not a task list. The public record is here: the 2024 Google Search documentation leak.

What gets sold locally that Google names as spam

Google publishes sixteen spam policies. Each is a documented way to be demoted or removed, and the risk lands on the business that owns the domain. Three describe products sold to small businesses in this area.

• Link spam. Google's definition is creating links to or from a site primarily to manipulate search rankings. That is the link-building package as commonly sold, described in Google's own words.
• Scaled content abuse. Pages generated primarily to manipulate rankings rather than to help people, and Google says this applies whether automation or humans produced them.
• Doorway abuse. Near-identical pages targeting similar queries, which is exactly what a proposal to build forty town pages produces.

The town-page product fails for a second reason unrelated to policy: those pages are sold to win map results, and pages do not rank in map results at all. The full list is published at Google's spam policies, and it is short enough to read before your next sales call.

One more mistake belongs here because it is the most expensive on the list: letting an agency own the domain, the hosting, the Business Profile or the ads account. The day that relationship ends, the business starts over.

What honest search work consists of for a business this size

Stripped of the monthly deck, the work that reliably matters for a local business is short and finishable: one clear page for each service, named the way a customer would say it out loud; the business's own words, including the towns actually served; a site that loads on a phone on a slow connection; titles, headings and internal links that describe the page honestly; a verified, correctly categorized Business Profile; and enough genuine reviews to clear the low double digits.

That list ends. It has a last item. Retainers exist in this industry largely because finishable work does not recur and a monthly invoice needs something to describe. There are real reasons to keep paying somebody, and they are worth naming: a site that changes often, a market with genuine competition, a store, or several locations. A five-page brochure site for a single-location trade business in Kaufman County is usually not one of them.

What drives the cost of the work worth doing is structural rather than mysterious: how many services need their own page, whether the existing site can be worked on or has to be rebuilt, and whether anything upstream is broken.

When you do not need this

You have one real competitor in town and more work than you can handle. If you are one of two septic companies serving Kemp and you are booked three weeks out, ranking first changes nothing about your revenue. Your constraint is capacity, not demand capture. The money belongs in a second truck, or in somebody who answers the phone after five.

Your customers are almost all referral and repeat. Search is a channel for people who do not already know you. If nobody arrives cold, you would be buying reach into a segment you do not serve.

Your market has essentially no search volume. Tool, Seven Points and Kemp are small enough that many trade searches there round to nothing in a month, and even Mabank and Gun Barrel City are thin for most categories. We would rather pull the real numbers for your trade than print an estimate here. If the whole market turns out to be a few dozen searches a month, an ongoing programme cannot pay for itself out of organic search alone, and we will say so.

Something upstream is broken. No hours on the profile, a voicemail box that is full, no photographs, a site from 2014 that fails on a phone. Search sends more people to a broken experience. Fixing the experience is cheaper and the effect is immediate.

You are still finding out whether the business works. Organic search has a payback period measured in months. If you need to know inside sixty days whether the offer sells, that is a paid advertising question, and the honest answer is often both at a small scale: advertising to make the phone ring now, and a handful of correct pages that keep working after the advertising stops.