DataLens for marketers 2026: dashboard without Looker and Power BI

Once a month, the founder writes to me with something like this: “We need BI, but Power BI is expensive, and Looker has fallen off.” In nine cases out of ten, this hides one task - to put together one dashboard for the marketer, where they can see traffic, conversions and CPL by channel in one place. You don't need Power BI or a paid subscription to do this. You need Yandex DataLens, and such a dashboard is assembled in an evening. I’ve been doing them for clients for three years now, and below I’ll discuss not “what BI is,” but specifically: what to connect, what charts to collect, and where DataLens has a fair ceiling.

Immediately frame. DataLens is not a Power BI killer or magic. This is a free dashboard designer from Yandex with native connectors to Metrica and Direct. For the RF stack, it covers 80% of the tasks of the marketing report. The remaining 20% ​​- complex calculations and heavy unloadings - hit the ceiling, about which the sellers of “turnkey end-to-end analytics setup” are silent. I’ll also tell you about the ceiling; without it, the article would be an advertisement.

A dashboard is not “pretty graphs”. A dashboard is one screen that you look at in the morning and within thirty seconds you understand where the money is going today.

1. What is DataLens and why is it free?

DataLens is a BI service in the Yandex ecosystem with Metrica, Direct and Cloud. The logic is simple: you connect a source, you build a dataset on it, you draw charts on the dataset, and you collect the charts in a dashboard. The same four layers as in Power BI and Looker Studio, only without a per-seat license.

“Free” is not a marketing trap with an asterisk. Connecting Metrica, Direct and Google Sheets, assembling charts, sharing - all this does not cost money and will not cost. You start paying when you go to Yandex Cloud with your database and big data: then the compute and storage are considered. But a marketer who reports on Direct and Metrica does not enter this zone. Over nine years in digital, I have seen hundreds of “free” tools that turned into paid ones on the second settings screen. DataLens is not one of them, and it is rare.

2. Connecting sources: what to pull from and where

The main advantage of DataLens for RuNet is its native connectors. Not uploads, not third-party parsers, but direct account connection. Here's what I connect to a typical project:

• Yandex Metrica - traffic, sources, behavior, goals. Authorization with an account, selection of a counter, fields are drawn on their own. This is also a convenient place to keep your tuned goals of Ya.Metrika as a basis for conversions.
• Yandex Direct - costs, clicks, impressions, CPC by campaign. Same login, choice of account. This is why DataLens is generally used for the Russian Federation.
• Google Sheets - expenses by channels that are not in the accounts: bloggers, sponsored placements, offline, manual estimates. The marketer keeps a sign, DataLens reads it.
• Database (ClickHouse, PostgreSQL, MySQL) - if you have a sales backend and need revenue, not just leads. This is where SQL or developer help comes in handy.

The Metrica plus Direct plus Google Sheets combination completes the basis. If you want to combine advertising with real sales from CRM, the source becomes a database or download, and this is the territory end-to-end analytics - a separate story, where DataLens plays the role of a showcase, and not a calculation engine.

One caveat about the linking of channels. In order for expenses from Google Sheets and sessions from Metrica to converge on one channel, a single markup is needed. Without neat UTM schemes in your dashboard you will have “yandex”, “Yandex”, “Yandex-direct” and “ydirect” as four different sources. This is not a DataLens bug, this is a mess in the input data. A pure UTM scheme saves you hours of fussing with calculation fields.

3. What charts do marketers really need?

This is where the main temptation of a newbie comes into play - to pile up twenty charts because “we can.” No need. A dashboard that no one looks at for more than five seconds is useless. The basic set that I put on the first sheet is six or seven blocks, no more.

Traffic dynamics by day or week. Funnel from visit to target action. Expenses and CPL by channel side by side so you can immediately see which channel is wasting your budget. Table of sources with CPC and conversion. And at the top there are a couple of large numbers - leads for the period and average CPL. All. The composition of specific charts and their sources are in the second table below.

The calculated fields for marketing in DataLens are simple. CPL is expenses divided by leads, conversion is actions divided by visits. The formula language is its own, not SQL or DAX, but anyone can understand the level of four arithmetic operations. For complex calculations like cohort retention or window functions, you will have to go to the documentation, and sometimes it turns out that it is easier to calculate in the source.

4. Dashboard assembly and selectors

Charts live on separate sheets, a dashboard is a sheet where you pull them out and arrange them. Layout logic from top to bottom: large numbers, then dynamics, then a breakdown by channel, then a detailed table. A person reads a dashboard like a page - from left to right, from top to bottom. If the most important thing is at the bottom, it won't be seen.

Selectors are what turns a picture into a tool. I put at least two: date period and channel. One selector filters all the charts at once, and the marketer himself says “show me only Direct for the last week” without new requests to the analyst. At EZ KATKA, where there were 19 clubs, the city selector saved me about an hour a week on manually cutting up reports - I simply gave the managers a dashboard with a filter and stopped being a spacer between them and the numbers.

5. Sharing and updating data

Access is granted in two ways. For accounts within the organization - when the team needs to look, but not outside. Public link - the dashboard opens without authorization at all, convenient for a client who does not have a Yandex account. The disadvantage of the public mode is severe: everyone who gets it will see the link. For revenue or margin figures, it is necessary to close public access, otherwise your data will be floating around in other people’s chats.

Data updating is not realtime, and thank God the marketer doesn’t need realtime. The Metrica and Direct connectors are pulled up when opened or on a schedule through materialization. Google Sheets is updated using the dataset cache: I set it from 15 minutes to an hour. A realtime dashboard is a different engineering task with a different budget, and in 99% of marketing reports it is superfluous. The picture “from yesterday” decides.

6. DataLens vs Looker Studio and Power BI

The main question on calls: “Why is it better than Looker?” The honest answer is that for the RF stack it’s better to use native connectors and something that won’t fall off. Since the summer of 2022, Looker Studio has not officially been friendly with Russian advertising sources, and Direct has to be pulled there through third-party paid connectors. Plus Google itself is unstable for Russian users. DataLens does not have these problems in principle.

Power BI is another league in terms of calculation power: the DAX language is more serious, the data model is more flexible, window functions are out of the box. But this is a Microsoft desktop stack, licensed, and there is no official support for Russia. If the task is a marketer’s report on Metrica and Direct, Power BI is like shooting sparrows out of a cannon. If the task is a million-row financial model with custom logic, DataLens will hit the ceiling, but Power BI will not. The case comparison is in the table.

7. Set of basic dashboard charts and sources

The specific composition of the very first sheet that I give to the client. Six blocks, each from a clear source. This is a skeleton - one or two charts are added under a niche, but the backbone is the same on any project.

If you want the same skeleton, but on the Google stack, I have a breakdown dashboards in Looker Studio. And I lay out the general logic of why a marketer needs analytics at all and in what sequence to build it, in an article about marketing analytics.

8. Where is DataLens' fair ceiling?

Now this is what they are silent about on selling landing pages. There are three places where DataLens stumbles.

Complex calculation fields. The language of the formulas is simple, and this is a double-edged sword: it calculates arithmetic easily, but cohort analysis, multi-stage attribution or non-trivial window functions are painful to write in it. I resisted a couple of times and sent the calculation to the source - in ClickHouse it’s ten lines, in DataLens it’s half a day of pain.

Large unloads. At hundreds of thousands of rows with a heavy data model, the dashboard starts to think. It can be treated by materialization and pre-aggregation in the database, but this is engineering, not “assembled in an evening.” If you have a million transactions and need flexible slicing across ten dimensions, DataLens will slow down.

Custom logic at the junction of several databases. When the model is a dozen related tables with different granularities, it is easier to put things in order in the storage and give DataLens a ready-made showcase. It is an excellent showcase and a mediocre calculation engine. Understanding this limit means not promising the client that the tool will not be taken away, and not disgracing yourself in a month.

Conclusion

DataLens is not a “free replacement for everything,” but an accurate tool for a specific task: a marketing dashboard for Metrica and Direct for a Russian project. In this niche, it beats both Looker Studio and Power BI - with native connectors, zero price and the fact that it will not fall off for political reasons. Where complex calculations and big data begin is where its place is on display, and the engine moves to the database. I have been collecting such dashboards for clients for three years now, and almost always the first working screen is born in one evening, and not in “a month of BI implementation.”

If your numbers are now spread across five tabs and an Excel file that no one opens, write to me in Telegram @dipustovalov or leave a request via form. Starting consultation - 0 ₽: in half an hour we’ll figure out what you have at the input, and we’ll tell you which dashboard will answer your question “where does the money go” in the first week.