Business and Service Websites: Using Visit Data to Identify Information Gaps during a Content Update

When updating business or service website content, relying solely on intuition can miss critical user needs. This guide explains how to leverage analytics data to pinpoint where information is lacking or outdated, ensuring your updates address real visitor questions rather than assumed ones.

Identifying Information Gaps Through Visit Data

As a brief hypothetical: imagine an author maintaining a small service company’s site and preparing a content update from visit data already on hand. The export shows three candidate pages by session volume and exit rate: “Pricing and What’s Included,” “Booking a Consultation,” and “Coverage and Response Times.” The first draws steady traffic but many visitors leave without opening any subpage; the second gets fewer visits but repeated returns to the same section; the third has low traffic yet the longest dwell time per visit. The author’s draft list pairs each with a possible gap: an unclear price range, uncertainty about what happens after booking, and missing detail on how quickly a request is handled in each service area.

From this, the author sketches working titles for a revised section on each page. One candidate is “What the Quote Covers and What It Doesn’t,” which would sit under the pricing page and try to answer objections that the current short paragraph does not address. A second is “After You Request a Consultation,” a short sequence of what the visitor should expect next, placed near the booking form. A third is “Typical Response Windows by Region,” a small table on the coverage page. None of these titles is treated as final, and none is drawn from real customer messages; they are self-made examples to show how a gap can be named before the content is rewritten.

One unresolved choice remains: the author cannot tell from visit counts alone whether the pricing page’s departures mean the price information is missing, or whether visitors find it and simply leave because the figure is higher than expected. The coverage page’s long dwell time could signal useful detail or confusion. Before committing to a rewrite, the author flags one item to check later—whether the booking page’s returning visitors are mostly using the same form or hunting for a confirmation message—and leaves the three titles side by side in the draft outline, not yet merged or assigned.

Interpreting Engagement Metrics and Returning Users

The dashboard’s engagement rate is the percentage of Google Analytics sessions meeting any of these conditions: a key event, a duration longer than 10 seconds, or two or more page views. Returning users is the percentage of users who started at least one previous session and came back; the source says this shows whether people return through organic Search.¹

If engagement rates drop, the source advises analyzing your content and checking whether it can be made more relevant to your organic search audience—for example, by ensuring page content is closely related to the queries people use to find you. Sessions and engagement rate belong together: sessions show search traffic volume, engagement rate its quality.¹

Before revising any page, you might pair each page’s engagement figures with its returning-user count and read the two together rather than separately. For example, the author could take the pricing page’s visit data already on hand: short stays plus a steady share of returning visitors may suggest those readers keep coming back without finding what they need, while short stays with almost no returning visitors may simply mean the page attracts one-time browsers.

To make this concrete, the author might sketch a small comparison beside the existing export: one column for average time on page, one for returning-visitor share, and one short note on what the page promises. For the pricing page, the note might read “quote coverage explained; no renewal terms.” Seeing the returning-visitor share stay flat while time on page drops could support adding a short renewal section, whereas a page whose returning readers stay longer may need only a clearer opening line.

You can choose which pages deserve this pairing first; a good candidate is any page where returning readers spend noticeably more time than new ones. If a section logs long visits and many returning readers but still shows exits near a single question, consider rewriting that question as a heading rather than adding more text. As a hypothetical revision, “What the Quote Covers and What It Doesn’t” might become “Quote Coverage, Exclusions, and Renewal Terms,” with the renewal sentence moved to the top. You might compare both titles against the visit pattern and keep the one that answers the repeated question sooner.

Prioritizing Pages Based on Traffic and Search Intent

The guidance cautions against focusing on traffic alone: more visits, more pageviews, or low bounce rates don’t necessarily equate to user success. Instead, it advises determining business goals, defining how success will be measured, and then gathering the right data to know when targets have been met. Keywords and search terms are listed as a way to surface the words people use to search for content and to identify topics they’re having trouble finding.²

For prioritization, the source suggests identifying and understanding an agency’s top-visited websites and web pages, noting that this data can help prioritize how to improve the user experience on the sites you manage. With the Digital Analytics Program (DAP) installed, the steps are to go to analytics.usa.gov, select your agency from the dropdown, select “30 Days” under the Top Web Pages and App Screens section, select “Download the full dataset,” and open the .csv file to analyze it in a spreadsheet. The source states the code should be installed at least 30 days ahead of when the report is needed.²

For each page in your visit export, you can compare where most entrances land with the words readers used to arrive. A page that draws many first-time readers from a broad search phrase is often the one where a short, plain answer near the top does the most good; a page reached through a narrower phrase may instead need one clear next step. If two pages pull similar entrance counts but only one shows readers leaving early, consider giving the livelier page’s opening treatment to the weaker one first.

As a hypothetical, an author might line up three pages: a services overview, a quote-request page, and a page explaining what the quote covers. The overview draws the most entrances from a wide phrase but readers leave quickly; the quote page draws fewer but holds them; the coverage page draws little. The author could choose the overview for the first revision and leave the coverage page alone, since little traffic offers little signal. That choice is yours to make, and a different site may reasonably start elsewhere.

You might also record, beside each candidate page, one question a reader likely brought and the sentence that would answer it. Writing that sentence first makes the priority concrete: pages where you can draft the answer in one line are usually quicker wins than pages that need research before any rewriting begins.

Conducting Targeted Content Audits with Analytics

Michelle Rago of 18F said she always includes metrics in an audit because they inform prioritization and can yield surprises, such as a page with incorrect information drawing heavy traffic as a top search result for a particular term. Ryan Johnson of GSA Service Delivery said that when he first did a content audit he left analytics data out, then had to merge in Digital Analytics Program (DAP) data afterward, when it could have been included during the audit.³

For other attributes, Rago captures HTML page titles and page headings and explores the differences between them, noting they are sometimes very different, not descriptive, or do not capture what the page is about. She also records purpose, audience, subject matter expert or content manager, goal, and a redundant, outdated, or trivial (ROT) analysis.³

On automation, Rago uses whatever tool she has access to, because automation can only get so far; at some point you need to read and experience the content, and a tool cannot know users' needs or organizational goals. Johnson uses automated tools for accessibility, SEO, reading grade level, and broken links, but still manually audits each page, and mostly uses automated scores to help prioritize next steps.³

Start with the visit export you already have and pair the quote-request page’s entry phrases with its exit counts; if several readers arrive asking about timing and leave without submitting, you can draft one explanatory sentence at the top of the form such as “Most quotes are prepared within two business days.” Read that sentence aloud against the existing opening lines, and keep whichever version answers the arrival question sooner.

You might then compare the services overview and the quote-request page side by side, using time on page and returning-visitor share as your two columns. As a hypothetical, suppose three returning readers lingered on the overview but only one reached the form; an author could soften the overview’s closing paragraph and add a single line pointing to the form, rather than writing a new page. This choice is yours to make after reviewing whether the delay reflects missing information or merely a long page.

For each revised page, you can note the question a reader likely brought and the sentence that now answers it, then check that sentence against the page’s original wording. If the two conflict, one sentence should go.

Aligning Updates with User Navigation Patterns

The Digital Analytics Program source advises establishing website goals and objectives before collecting data, then choosing metrics tied to defined calls to action and key performance indicators rather than focusing on traffic alone. It notes that web analytics, clickthroughs, scroll-tracking, and heatmaps show how people navigate, and that keywords and search terms surface the words people use and reveal topics they are having trouble finding.²

The source also recommends identifying and understanding top-visited websites and web pages, stating that this data can help prioritize user-experience improvements. Separately, it cautions that more visits, more pageviews, or low bounce rates do not necessarily equate to user success.²

Begin with the visit export in front of you and pick one page whose entry and exit patterns don’t match the page’s promise. For example, the quote-request page may draw readers who arrive on a query about timing, then leave quickly. You might write the phrase groups from that query beside the exit count and ask whether an early sentence addresses timing before the request form begins.

Consider reordering rather than rewriting first. As a hypothetical, take a draft line such as “Tell us about your project and we’ll respond,” and place a sentence about expected response time above it. This is an author’s illustrative edit, not a tested improvement. Your own export can tell you which page deserves this kind of small adjustment; you can choose one where entering phrases and departing readers point to the same unanswered question.

Then read the revised sentence against the page’s original wording and the surrounding section, checking that nothing now overpromises. If the revised page mixes services and pricing, you might move the timing sentence to whichever section readers reach first. Keep a short note of the question you matched and the sentence you changed, so a later update can revisit that choice without reopening every draft.

Establishing Continuous Maintenance Processes

The source frames maintenance as an ongoing process rather than a one-time edit: it recommends establishing processes to maintain quality and accuracy, setting up a governance structure, and standing up a content team. It also calls for continually assessing your digital presence so content stays the right size for your audience's needs.⁴

It describes web analytics as a way to track the views a page gets, noting that a page receiving little to no views might be ready to retire, and content audits as a way to decide when to retire, consolidate, or update content that is accessed frequently. It also states that governance strategies help determine the right metrics and assign ownership, including verifying content with subject matter experts, and that small teams can set formal roles and responsibilities and publish an initial minimum viable product before adding enhancements over time.⁴

Set a recurring review rhythm you can actually keep: choose a monthly or quarterly slot, open the most recent visit export, and compare it with the notes you kept from the previous pass. You can pick just one page per cycle, so the work stays small; the quote-request page is a reasonable first candidate if its entry phrases and exits still disagree.

At each pass, you might reread your earlier sentence—the one meant to answer the reader’s likely question—beside the page’s current wording. As a hypothetical, a draft line such as “Tell us about your project and we’ll respond” could gain the sentence “Most quotes are returned within two working days,” if your own notes suggest timing is the recurring question. You can mark whether the new sentence still matches the surrounding section and whether it now overpromises.

Record one change per cycle in your own file: the page, the question, and the sentence you revised. Consider leaving a short note on what you will check next time, such as whether returning readers stay longer. You decide the cadence and whether to pause a page; the point is a light, repeatable habit rather than a completed overhaul.

Frequently asked questions

How might I combine a page's engagement figures with its returning-user count before revising it?

Before revising any page, you might pair each page’s engagement figures with its returning-user count and read the two together rather than separately.

What could I record beside each candidate page to organize my notes?

You might also record, beside each candidate page, one question a reader likely brought and the sentence that would answer it.

What should I record in my own file for each change cycle?

Record one change per cycle in your own file: the page, the question, and the sentence you revised. Consider leaving a short note on what you will check next time, such as whether returning readers stay longer.