SEO Has Produced No Results After Six Months—Should You Keep Publishing Articles? When Should You Adjust the Original Strategy?

The IKEA effect refers to the tendency for people to place a higher value on things they helped create, because those things contain their own time and effort. In SEO, this turns into a state that is hard to self-audit: the team has already written for six months, held more than a dozen topic-planning meetings, and revised the template three times, so “we are doing SEO” itself becomes an achievement, while “has this batch of content brought in inquiries” becomes an awkward question to ask.

It usually has three observable manifestations in a project: treating output as progress, with weekly reports only listing the number of posts published; being unwilling to delete pages even if some pages have not received a single click for a year; and treating “wait another quarter” as the default option without ever defining what “success” would look like when the wait is over.

Acknowledging this is not a denial of effort; it is handing evaluation authority back to the data. The three scenarios below are typical forms of high investment and low return.

Three Typical Scenarios: High Investment, Low Growth

The Content Hub Published Hundreds of Articles but Did Not Bring in Qualified Customers

The number of articles and indexed pages keeps rising, but form submissions, inquiries, and trial requests barely move. The reason is often not “not writing enough,” but that topics revolve around industry news instead of purchasing questions: users search for “how to do it, how much it costs, which solution fits me,” while the content answers “what happened in this industry recently.” Another common problem is that every article ends with a vague mention of the product without giving a next step.

The way to handle this is to pull the articles from the past 90 days that have impressions but no conversions and check the intent of each one—whether the searcher is in the awareness, comparison, or decision stage, and whether the page has the corresponding answer and entry point. The priority for revising these pages is higher than writing new ones.

Technical SEO Consumed a Large Amount of Development Resources Without a Clear Ranking Improvement

Rebuilding URLs, changing templates, adding structured data, and optimizing Core Web Vitals are not wrong in themselves, but technical fixes can only give “better content a chance to be seen by the right users”; they cannot replace the content itself. The benefit is conditional: after the foundation is cleaned up, content that was weak remains weak.

The way to verify this is to record the average ranking and click-through rate of a group of target pages before the technical launch, and compare them 4 to 8 weeks after launch. If crawling, indexing, and page experience all improve but rankings do not move, the problem is not at the technical layer; do not keep adding development resources.

A Previously Effective Keyword Strategy Stopped Growing When Expanded Further

The first batch of keywords once brought leads, so the team used the same template to expand to two or three hundred long-tail terms. The result was rising impressions, falling click-through rates, and declining lead quality. This usually means the effective radius of the original strategy has been used up: there is only so much demand with real commercial value, and horizontal expansion merely dilutes site-wide authority and content quality.

The way to handle this is to review in three layers—“keyword—page—lead”—identify the page types that have actually generated leads, and pull resources back to go deeper instead of continuing to spread wider.

How Long Without SEO Results Before Reevaluation Is Needed?

No official timetable can answer this question; industry competition, site history, and existing content vary too much. An operational approach is to set an observation window and a “minimum signal” for each type of work, and enter reevaluation when the window expires without a signal. The table below is an internal management baseline for triggering reevaluation, not a promise from search engines and not a ranking guarantee.

Execution Problem or Direction Problem: A Four-Layer Diagnosis

Check from the bottom up across four layers. Stop at the layer where the chain breaks; do not jump to conclusions.

1. Crawling and indexing. Use the sitemap and index coverage report to confirm that target pages have been indexed; if not indexed, first solve crawlability, canonicalization, and duplicate content.
2. Impressions. In the Performance report in Search Console, check the number of impressions for queries related to the topic. If impressions are zero, the page has not entered the candidate pool; the problem is demand, intent, or page authority, not copy wording.
3. Clicks and rankings. If there are impressions and the average ranking is outside the top 10, demand exists but competitiveness is insufficient; prioritize adding content depth and improving page experience. If there are impressions, rankings are high, but click-through rate is clearly lower than for other queries on the page, the title and snippet do not match intent.
4. Conversion. If there are clicks and time on page but no conversion, the problem is usually consistency between the landing page’s promise, evidence density, and the next-step entry point. This is a page problem, not an SEO problem.

Google Search Central has a separate document on troubleshooting traffic drops, and the methodology is consistent with this: first confirm the scope of impact (site-wide, directory-level, or single page), then align the timing (redesign, migration, algorithm update), and finally confirm whether the data is a short-term fluctuation. For the specific process, see Debugging Google Search traffic drops.

Decision Table: What Action Matches the Current Situation

To turn diagnostic conclusions into action, you can use this table directly. The left side is the current situation visible in search data; the right side is the recommended action.

How to Set Stage Goals: Replace “How Many Articles to Publish” with “What to Validate”

Replace “publish 60 articles this quarter” with “validate three hypotheses this quarter,” and SEO becomes manageable. A usable quarterly goal contains three things: a hypothesis, a metric, and a stop-loss line.

Take content direction validation as an example:

• Hypothesis: Mid-sized and large customers will search for questions such as “solution comparison, pricing, and implementation timeline” during the selection stage, and there are no corresponding pages on the site.
• Metric: Within 8 weeks, the topic brings observable impressions and clicks, and some of those visits enter the pricing or contact page.
• Stop-loss line: If it still has not entered the top 30 in average ranking after 12 weeks, stop expanding and instead add sections to existing pages with high impressions.

Two details determine whether this method works: metrics must come from search and site data, not subjective judgment; and the stop-loss line must be written down before launch—adding it afterward is the same as not having one.

Which Pages Should Be Kept, and Which Should Be Merged or Updated

• Keep: Core pages that have received clicks in the past six months, have backlinks, have queries ranking in the top 20, or serve brand and product definition functions.
• Merge: A group of pages whose topics overlap heavily, cannibalize each other’s rankings, and have insufficient information when viewed individually. Choose the one with the best data as the main page, merge the useful content from the other pages into it, and then redirect them.
• Update: Pages with impressions, rankings between 10 and 30, and still-existing demand, but with outdated information or missing evidence. Prioritize updating them; the cost is far lower than writing new ones.
• Discontinue: Pages that have had no impressions for a long time, have no internal links, and do not carry brand information. Stop maintaining them and evaluate taking them offline.

Execution steps:

1. Export page-level clicks and impressions from Search Console and cluster them by topic.
2. Tag each page with a role: definition, comparison, tutorial, case study, tool.
3. Within the page group for the same role and same topic, select the main page.
4. After merging content, set redirects or canonical URLs, and update internal links and the sitemap for the main page.
5. After 4 to 8 weeks, look at the overall impressions and rankings of the merged topic, not the drop of a single URL.

Which SEO Projects Should Stop Receiving Further Investment

The following types of projects are worth putting directly on the stop list:

• Industry news produced to make up the numbers, with no next step for readers after they finish.
• Long-tail keyword matrices that are already outside the top 30, have no backlinks, and have never been clicked, yet continue to be expanded.
• Template-level optimization that repeatedly tweaks styles but never changes the page’s information architecture.
• Topic selection that serves internal reporting keyword coverage rather than real search demand.
• Backlink or content partnerships that are expensive to maintain and unrelated to the core business.

Stopping does not mean giving up entirely. Concentrating the saved budget on a small number of pages that have impressions, are close to the first page, and are related to leads is usually more effective than opening ten new directions at the same time.

In the AI Search Era, Evaluation Needs One More Layer

In addition to rankings and clicks, you should also look at one thing: whether your pages can be cited by AI search and Q&A products.

This is not mysterious. The direction given by Google Search Central is still that content should be genuinely helpful to people, information should be verifiable, and structure should be clear (see Google Search Central and Creating helpful, reliable, people-first content). On this basis, pages that are easy to cite usually share several traits: the opening gives the conclusion directly, there are judgments and data that can be excerpted independently, entities and terminology are expressed consistently, and there is a stable update time and author information.

A simple self-check: excerpt the conclusion on the page that you most want to be cited and read it on its own. If it is a complete answer that stands without context, it has a chance of being cited; if it is all buildup and adjectives, rewrite that paragraph first.

30-Day Review: In What Order to Do It

If you decide to conduct a formal review, you can complete it within 30 days in this order.

• Week 1, freeze the data. Export page-level clicks, impressions, average rankings, and conversion paths for the past 12 months, and tag them by topic.
• Week 2, diagnose by layer. Locate breakpoints across the four layers—crawling/indexing, impressions, clicks, and conversion—and write a conclusion for each topic.
• Week 3, make trade-offs. Use the decision table to produce four lists—keep, merge, update, and discontinue—and estimate the manpower required for each type of action.
• Week 4, reset goals. Write down the hypotheses, metrics, and stop-loss lines for the next quarter, and determine the publishing cadence at the same time—the focus is no longer “how many to publish,” but “what to validate.”

After that, do only two things each month: advance the lists and compare intermediate metrics.

Back to the Original Question

Whether to keep publishing articles depends on whether you can say which pages generated impressions during the past six months, which queries are close to the first page, and which content brought leads. If you can say this clearly, you should continue, but change the cadence and focus. If you cannot, you should stop and review first, rather than adding another batch of articles.

If the bottleneck is the website itself—the page structure does not support layered information, every article has to wait for development scheduling, and changing a title takes a three-day process—then the problem is not content output but the website that carries the content. An AI website-building platform for showcase sites such as We0 puts the AI website generator, multi-agent collaboration, CMS backend, SEO and GEO optimization, and domain deployment on the same chain, so teams can adjust page structures and content templates themselves and validate the hypotheses above with shorter iteration cycles (We0 website). Tools cannot decide strategy for you, but they can make “try changing it once” cheap enough.

FAQ

If SEO has had no effect after six months, should all content updates be stopped immediately?
A blanket stop is not recommended. First use the four-layer diagnosis to confirm the breakpoint: if indexing and impressions are both growing and only conversion is not moving, the problem is the page rather than output, and stopping updates at this point would interrupt momentum that has already been validated; if there are almost no impressions at all, you should stop horizontal expansion and shift resources to fixing pages and changing topic selection.

How many articles per month count as acceptable?
The number of articles is not the metric. A manageable goal is “what hypothesis was validated this period, and what intermediate signals appeared within the observation window.” Under the same budget, improving ten pages that already have impressions is often closer to leads than writing thirty new topics with no query baseline.

Will merging old articles lose existing rankings?
Merging targets page groups with overlapping topics that compete for rankings. The approach is to keep the main page with the best data, merge the useful content from the other pages into it, redirect them, and update internal links and the sitemap. When evaluating, look at changes in impressions and rankings for the whole topic, not the drop of a single URL.

How do you determine that a keyword direction is no longer worth further investment?
Look at three points: no impressions by the end of the observation window, no natural backlinks to the page, and no historical visits from that topic have ever entered the conversion path. If all three are true at the same time, you can stop expanding and move the budget to topics that are already close to the first page.

If you are not cited in AI search, was the SEO work wasted?
Both share the same content foundation. The prerequisite for being cited is usually that the page has a clear conclusion, verifiable information, and consistent wording—exactly the capabilities rankings also require. Fix answerability first, then discuss visibility in AI channels; this is more stable than chasing a single channel in isolation.

Related Tools

• Google Search Console: View query-level impressions, clicks, index coverage, and page experience data—the foundational data source for the four-layer diagnosis.
• Google Analytics 4: Connect organic search traffic to conversion events such as forms, sign-ups, and inquiries.
• Sitemap and canonical checks: Confirm that target pages are correctly indexed and that duplicate URLs are not cannibalizing each other.
• Content inventory table: Cluster pages by topic and role to support decisions to keep, merge, update, or discontinue.

Related Links

• Google Search Central (Chinese)
• Debugging Google Search traffic drops
• Creating Helpful, Reliable, People-First Content
• We0 website

Summary

SEO not working after six months does not mean SEO is ineffective; it is more like a signal that a review is due. Use a four-layer diagnosis to distinguish execution problems from direction problems, use observation windows and minimum signals to decide whether to continue or adjust, use a decision table to handle five high-frequency situations—“impressions but no rankings,” “rankings but no conversions,” “duplicate content,” “technical work with no return,” and “no real demand”—and then use stage goals to tie resources to testable hypotheses. What truly needs to stop is never SEO itself, but those projects that have already proven to have no commercial value yet are hard to let go of because too much has been invested.