Analytics Without Goals: Why the Dashboard Changes Nothing on Monday

The dashboard is full and the Monday meeting is unchanged

The usual starting point is not an absence of analytics. There is a counter on the site, a chart that rises and falls, and a monthly export that one person forwards to another. What is missing is an answer to one question: which number, if it moved, would make you do something different this week? Without that answer the analytics is decoration. It records the business instead of steering it, and everybody in the room quietly learns to scroll past it.

The catalogue line that addresses this is Setting up analytics, goals and summary dashboard, filed under Analytics at $150 – $600, with a duration of 1–2 days and a difficulty of Beginner. In roubles the same line reads 5 000 – 25 000 ₽, with the catalogue noting that a basic market setup sits at 5 000–15 000 ₽ and that merging sources into a dashboard takes it up to 25 000. The description ends on the outcome rather than the tooling: the client stops making decisions based on feelings.

Read the price band as a description of scope, not of quality. The lower end buys correctly marked events and goals on a single site. The upper end buys the same work plus what the catalogue calls compiling data from all systems into one table. Both ends produce numbers. Only the second removes the argument about whose numbers are correct, and that argument is what usually stops a decision from ever being taken.

An event is a fact; a goal is a promise about what you will do with it

Marking events is mechanical work. A form is submitted, a call button is pressed, a cart reaches checkout, a price list is downloaded — each becomes a named record with a timestamp attached. This part deserves the Beginner label the catalogue puts on the job. It does not need a strategist. It needs somebody willing to sit down and list every action a stranger can perform on the site, then give each one a name that still makes sense to whoever inherits the account.

A goal is a different kind of object. A goal declares which of those events counts as progress, what volume of it is normal, and who is accountable when the number falls. The catalogue phrases the result as finally seeing what exactly brings applications. That sentence hides a commitment: somebody has to be willing to spend less on the channels that bring nothing, and more on the ones that do, once the table says so.

This is where most installations quietly fail. Events get marked, a dashboard gets built, and nobody writes down the threshold that triggers an action. A goal without a threshold is a counter with ambition. If the honest answer to what will we change if this halves is nothing, the useful move is to delete the metric rather than keep reporting it every month to people who cannot act on it.

There is a simple drafting exercise that separates the two. For each goal, finish the sentence: if this number stays below X for the agreed window, then Y happens and Z signs it off. Goals that cannot be finished in that form are not goals, they are curiosities. Curiosities are fine to keep on a second tab, but they should never occupy the top row of a dashboard that a busy person opens between two calls.

One table instead of four browser tabs open at once

The practical reason decisions stall is rarely ignorance. It is contradiction. The ad platform reports one number of conversions, the CRM reports another, the site counter reports a third, and the sales manager remembers a fourth. The meeting then opens by deciding which source to believe today and closes by agreeing to look again next month. Nothing has been decided, but everybody has been busy, and the calendar shows a meeting that happened.

The catalogue answers this in one clause: data from all systems is compiled into one table and reports are collected themselves. One table is not a cosmetic improvement. It forces the definitions to be reconciled once, in writing, at setup time, instead of being renegotiated informally in every meeting from here on by people who each have a tab open on a different service and no reason to yield.

The duration is the part worth staring at: 1–2 days. Whatever is stopping a company from having this, it is not a two-day technical project. It is that nobody has been asked to decide what counts as a lead, which sources are allowed into the table, and who reads it on Monday morning. The catalogue prices the technical half of the job; the definitions remain the buyer's homework, and they are due before the work starts.

The weekly rank report earns its place by ending with what to do

The second line is Monitoring of positions with automatic report every week, in the Monitoring category, difficulty Beginner. Its description sets out four things the report contains: what has grown, what has fallen, why, and what to do about it. It also states what disappears — no manual uploads and no logins to four services. Every Monday the report arrives whether or not anyone remembered to ask for it, which is the entire content of the word automatic.

The commercial terms are stated in two parts. Setup is $150–400, and in roubles 10 000–25 000 ₽, followed by a subscription. The duration line separates the two phases just as plainly: 1 day to connect a client, then 0–30 minutes per week. A buyer should read that as one short project plus a recurring obligation, and should settle what the subscription covers, and for how many queries, before signing anything.

The fourth clause of the description carries all the weight. A report that stops at what has grown and what has fallen is data. A report that ends with what to do about it is a draft decision, and a draft decision can be approved, rejected or amended in a single short conversation. That is the whole difference this article is about, printed inside a single catalogue description that costs $150–400 to switch on.

Reading 0–30 minutes per week as a staffing line, not a feature

The duration field for weekly monitoring reads 1 day to connect a client, then 0–30 minutes per week. That band is unusual because it includes zero. Zero means some weeks the report says nothing that requires a human at all. Thirty minutes means the week where something genuinely moved and somebody has to decide whether it matters, which is precisely the week you want a person present rather than another chart.

Here is an arithmetic example a reader can do with the catalogue's own figures, and it is an example rather than a measurement of any client: at the top of the published band, 30 minutes a week is about two hours a month and roughly 26 hours a year. Put that beside whatever your own team currently spends assembling the same picture by hand, and the comparison quietly turns into a staffing question instead of a software question.

Nothing in the catalogue claims that your current manual routine costs more than that. It claims only that the manual routine disappears — no manual uploads, no logins to four services. You are the only person who can price what it costs today, and it is worth timing it once, honestly, with a stopwatch, before you buy anything. That figure is the one a finance director will ask for, and the one nobody ever has ready.

A one-star review is an operations event with a clock on it

The third line is Review collector: all sites in one place with an alert for negativity, filed under Reputation with a difficulty of Experienced. The description names the sources — Yandex.Maps, 2GIS, Otzovik and marketplaces — and states the rule that makes the thing operational: new reviews are collected in one table, and negative ones, 1-3 stars, go instantly to Telegram to the person in charge. The stated result is that a bad review is closed in an hour, not in a week.

Notice that the decision has been made in advance and encoded. Someone has already decided that one to three stars is the boundary, that crossing the boundary triggers a message rather than a row in a monthly report, and that the message has a named recipient. A dashboard would have shown the same review a week later, in aggregate, next to a rating average that had moved by a rounding error and prompted nobody.

The published cost is $150 - $450, or 10 000 - 35 000 ₽, with a duration of 1-2 days for 3 sites. That last phrase is a unit, not a total. Three platforms is the quoted scope, so a business listed on six should expect the conversation about scope to come before the conversation about price, and should decide in advance which platforms genuinely influence a buyer and which are simply present.

The review collector is the only Experienced job of the three

Two of these three lines carry the difficulty Beginner: the analytics and goals setup, and the weekly position monitoring. The review collector carries Experienced. The catalogue is not grading the buyer's intelligence; it is signalling where the work is fragile. Reading a search ranking is a stable, well-defined operation with a small number of outcomes. Reading a review page is not, because every platform presents, paginates and protects its content in its own way.

That fragility is why the duration is expressed per platform — 1-2 days for 3 sites — rather than as a single flat figure. Each additional source is a new format to parse, a new failure mode to detect, and a new definition of what counts as a fresh review rather than one already seen. The alert rule stays identical across all of them; the plumbing underneath does not, and the plumbing is what consumes the days.

For a buyer this has one practical consequence worth writing into the brief. Ask which platforms are included by name in the quote, and ask what happens operationally when one of them changes its layout. A collector that silently stops collecting is worse than no collector at all, because the silence reads exactly like good news right up until a customer mentions the review you never saw.

Goals first, ranks second, reviews third

If the three are bought in sequence rather than together, the order that makes sense starts with analytics and goals. Ranks and reviews both produce numbers whose importance can only be judged against a goal. A ranking that fell means nothing until you know whether that query ever produced an application, and a run of negative reviews means something very different for a business that fields a handful of enquiries a week than for one that fields hundreds. The reverse order is the one bought more often, because an alert that makes a phone buzz feels like control long before anyone can say whether the reviews were costing money.

Adding the published figures gives a simple frame, and this is addition of catalogue numbers rather than a quotation from it. Taking the low end of each setup — $150 for analytics and goals, $150 for monitoring setup, $150 for the review collector — the three setups begin at $450 together. Taking the high ends, $600, $400 and $450, they reach $1,450, with the monitoring subscription still to be agreed separately on top of that.

The time side adds up just as plainly: 1–2 days for analytics and goals, 1 day to connect weekly monitoring, and 1-2 days for a three-platform review collector. That is between three and five working days of setup for an entire measurement layer. The calendar risk was never the work. It is the stretch of calendar that passes while somebody inside the company decides what counts as a lead and who owns the answer.

Before paying for any of it, name whose Monday changes

Every one of these three services ends at a person. The dashboard has a reader. The Monday report has an owner who can approve or reject the recommendation it ends with. The negative-review alert has, in the catalogue's own words, a person in charge. If you cannot name those three people before you buy, the purchase will reliably produce data on schedule and decisions never, and the renewal conversation will be an uncomfortable one.

The test is cheap and can be run this afternoon. Write down the three names. Then write, beside each one, the single action that person is permitted to take without asking anyone else — pause a campaign, rewrite a page, publish a reply to a review. If the second column is empty, fix the second column first. The software will still be there next week at the same published price, and it will work better when it arrives.

This is what the difference between data and a decision actually looks like in a purchase order. The catalogue can sell marked events, one consolidated table, a Monday report that ends with a recommendation, and an instant alert on a one-star review — at $150 – $600, $150–400 and $150 - $450 respectively. What it cannot sell is the sentence that says what happens next. That sentence costs nothing and is skipped more often than any other part of the system.