Why You Cannot Properly Improve a Website Without Analytics

Analytics feels optional only when teams still believe traffic growth can be managed by intuition and surface-level counters.

In practice, every serious growth decision depends on measurement: what channel created the lead, where the page leaked users, and which fixes moved revenue.

This article explains why analytics is not a reporting layer added later, but part of the operating system of the website itself.

Why “site without analytics” = “site without feedback”

Analytics feels optional only when teams still believe traffic growth can be managed by intuition and surface-level counters.

In practice, every serious growth decision depends on measurement: what channel created the lead, where the page leaked users, and which fixes moved revenue.

This article explains why analytics is not a reporting layer added later, but part of the operating system of the website itself.

Four layers of normal analytical infrastructure

In order not to confuse “we have a counter” with “we have analytics set up”, it is useful to take one look at the four-layer model. These are not stages in time and not a sequence of actions - these are parallel layers, each of which provides its own type of information for decision making.

Key Point: These Layers Workjointly and cumulatively. By itself, layer 1 (counter) shows abstract numbers without business meaning. Layer 2 without layer 3 is “interesting observations” with no connection to revenue. Layer 3 without layer 4 are events that are not associated with actual transactions. And vice versa - layer 4 without layers 1–3 is technically impossible to configure at all. Only a combination of all four layers gives a full effect, and this is what needs to be kept in mind when they say “customized analytics.”

Eight signs your analytics are broken

Below is a practical set of symptoms, each of which signals that analytics formally “exists”, but actually does not work. If you find out two or three or more, this is the very case when further promotion and improvement of the site is limited by the lack of feedback.

“Analytics out of the box” vs properly configured analytics

The difference between “installed a counter” and “set up analytics” is not quantitative, but qualitative. Below is a parametric comparison of the two states.

The difference between the columns is the difference between “advertising and SEO as a cost without the ability to measure the result” and “advertising and SEO as tools with a measurable ROI.” The first state is what most companies actually have when they say “we have analytics.” The second thing is what you need to come to before scaling up any marketing expenses.

Six myths about analytics that prevent people from setting it up

The question “is full-fledged analytics necessary” usually rests on several persistent myths. Below are the six most common ones and why they prevent businesses from making the right decisions.

What specifically cannot be done without analytics

Often the question “do we need analytics” is perceived in the abstract. But if you go through the specific tasks of website promotion and improvement, it becomes clear that without analytics, most of them are simply technically impossible. Below is a set of practical situations.

The second class of tasks that are impossible without analytics is the prioritization of website improvements. At any moment, the team has 10–50 ideas for “what could be improved”: recolor the button, shorten the form, rewrite the first screen, add a review block. Without data on where users are being lost, these ideas are ranked based on taste and team voice rather than actual business impact. As a result, they rule over what “catches the eye”, and not what really costs money.

Third class isunderstanding which marketing channel actually leads to transactions. Without configured attribution and connection with CRM, all channels look equally suspicious: “where did these 50 requests come from last month” is an unanswered question. As a result, the business evenly distributes the budget between channels, some of which work and some of which do not. With the right analytics, the picture becomes clearer: you can see which Google Ads really lead to money and which waste clicks; which SEO keywords convert and which only attract traffic; which mailings work and which ones simply create failures.^[4]

Formula for losses from blind promotion in BYN

To make the conversation about the need for analytics substantive, let’s try to calculate how much its absence actually costs. The formula is simple, but it shows the scale of the losses.

A typical scenario for an average commercial website. Advertising budget - 15,000 BYN/month. Without analytics, 30–40% goes to ineffective campaigns = ~5,000 BYN/month wasted. Lost revenue from correct prioritization of improvements (with a site CR of 1.2% instead of a potential 2.0% with focused work) - ~40,000 BYN/month of lost revenue. CPA growth due to stupid Google Ads strategies (+30–60% to CPA) - ~8,000 BYN/month overpayment. Total:~53,000 BYN/month, or~636 thousand BYN/year. And “setting up normal analytics” means 10–20 hours of specialist work, that is, ~3,000–6,000 BYN one-time costs. The gap is ~100 times.

To this should be added “invisible” losses: lack of alerts for critical events (for example, the form on the site does not work for weeks and no one knows), the inability to prove to the contractor that “advertising does not pay off” (without numbers it is always “yes, everything works for us”), growing dependence on someone else’s interpretation of data. All these are not “additional costs”, but direct damage to the business, which analytics automatically prevents.

Four areas of work with analytics

To understand where to start and where to move, it is useful to divide your work with analytics into four parallel directions. Each gives its own return and requires different qualifications.

Analytics settings checklist - 15 points

Below is a step-by-step checklist, according to which you can either set up analytics from scratch or diagnose an existing one. Each item is a clear check mark and an artifact.

How Ontop solves this

In our projects, analytics is a mandatory part of launching a website and advertising campaigns, and not “an additional service if there is time.” We do not launch advertising without correctly configured conversions and do not consider the site ready if it does not have at least a combination of GA4 + Metrica + GTM with configured events. Practice shows: “let’s launch it, then set up analytics” almost always turns into wasted weeks of advertising budget without the possibility of optimization.

Technically, our standard stack is a combination of Google Tag Manager + GA4 + Yandex Metrica; all tags (Google Ads Conversion, Yandex Direct, social network pixels, Hotjar, if necessary) are managed through GTM. Events are configured using a single method: each key action (submitting a form, clicking on a phone number, going to the messenger, scrolling to a form, adding to cart) is both a GA4 event, a Metrics goal, and a conversion in Ads. Through Enhanced Conversions and Conversion API, conversions are enhanced by CRM data - so that Google Ads sees not only leads from the site, but also real transactions with their amount.

We pay special attentionlink “analytics → advertising campaigns”. For each client, UTM markup of all sources is configured, GCLID/YCLID are saved in CRM, and offline conversions from CRM are transferred back to Ads and Direct. This allows the algorithms to learn from real results (deals and their amounts), and not just from “formal” leads. As a result, Google smart strategies and Yandex Autostrategies begin to work many times more efficiently, andadvertising is starting to pay offin those niches where it would not formally “work” without this.

Finally, we build automatic dashboards for each client in Looker Studio - a single screen on which in 5 seconds a picture of all key metrics is visible: traffic by source, CR by channel, ROAS, funnel from session to deal. This is the moment when analytics moves from a “technical responsibility” to a “daily management tool.” This is how our whole scheme works -single full-service agency, where is the development,SEO strategyand advertising is not three separate contracts, but one interconnected system with common analytics and common KPIs.

Frequently asked questions

Conclusion

A website without analytics can still exist online, but it cannot improve with discipline.

Once growth depends on prioritisation, attribution, and conversion economics, analytics stops being a dashboard and becomes infrastructure.

The economics of this solution are obvious. Without analytics, a site with an advertising budget of 15,000 BYN/month loses about 50,000 BYN/month - due to ineffective campaigns, incorrect prioritization of improvements, and increased CPA due to untrained strategies. The cost of setting up analytics is less than one week's loss. This is one of the fastest return on investment in a website available to businesses in 2026. The only way not to do it is to agree to blindness, and it is the business that pays the price of this blindness, and not the “contractor who looks at the reports.”

Sources

1. Google Analytics 4 - Introduction and setup— Google's official guide to launching GA4: creating a resource, setting up a data stream, basic configuration, communication with Google Ads and Search Console. Entry point into the Google analytics ecosystem.
2. Google Analytics 4 - Events, conversions and measurement— documentation on events and conversions in GA4: auto-collected and custom events, setting up key events (conversions), parameters, real examples for different types of sites.
3. Google Ads - About conversion tracking— the official Google Ads guide to conversion tracking: why it is critical for Smart Bidding, what types of conversions there are, how to set them up correctly to optimize advertising campaigns based on real user actions.
4. Google Ads - Enhanced Conversions— Enhanced Conversions documentation: transferring hashed user data along with conversions, increasing attribution accuracy by 10–20%, privacy policy compatibility, practical setup steps for different stacks.
5. Google Tag Manager - Developer documentation— Google's official guide to GTM: principles of centralized tag management, setting up triggers and variables, working with Data Layer, integrating GA4, Ads, social network pixels and third-party tools.
6. Yandex Metrica - Help— official documentation of Yandex Metrics: installation, configuration of goals and events, Webvisor (session recording), click maps, segmentation, integration with Yandex Direct and Webmaster. A key analytical tool for the Russian-speaking market.