How to calculate the ROI of an entire project: end-to-end analytics in simple words. What does end-to-end web analytics mean?

Without end-to-end analytics, all your efforts to make money online can turn into one endless nightmare. Advertising doesn't work - and you don't know why. People don't buy - and you don't understand how to fix it. The site doesn't sell - and there's nothing you can do about it.

To prevent this from happening, you need an “X-ray” that will show exactly what is happening in your funnels and how to improve them. End-to-end analytics will become such an X-ray.

In this article we will look at the simplest and effective scheme Setting up end-to-end analytics in 5 steps. And you can do it for free, with your own hands, literally in one evening.

As a short introduction, let's look at the example of my website to see how end-to-end analytics systems work and what they are needed for.

What is end-to-end analytics and how it works

End-to-end analytics is a system that allows you to track everything that happens inside your sales funnel. Starting from the moment a potential client first sees your ad - right up to when he places an order and pays money.

Sometimes all these actions can be tracked in one place, and sometimes you need to configure several programs to do this. But the main thing is that all your indicators are before your eyes. and you can control them. And when we control something, it means we can manage it (increase and improve).

Here, for example, is what a sales funnel with configured end-to-end analytics might look like.

An example of end-to-end analytics in action

In order not to go far, let's look at this site. whichever you are on right now. From the outside it may not be noticeable, but I also have my own sales funnel with configured end-to-end analytics.

I warn you right away that my sales funnel is built a little unusually, because this is blogging, information business, and other “obscurantism”. In a more traditional business it will look a little different.

But we will look at and build a traditional funnel a little lower, and here we at least have the opportunity to look at specific numbers.

So, here are the steps my sales funnel consists of:

1. A person goes to my site
2. A person goes to the subscription page for my “Contactless sales” newsletter
3. A person subscribes to my newsletter
4. A person studies my free materials and purchases a paid package

Everything is so simple and clear. But to track all the indicators, I have to use two whole various systems analysts.

The first system is tracking a person from the moment they visit the site until the moment they subscribe (step 1 - 3). Here I use Yandex Metrica. The second system is tracking a person from the moment of subscription to the moment of purchase (steps 3 - 4). Here I am using the service JustClick.

How I find weak points in my funnel

To understand why I haven’t become a billionaire yet, I open the indicators of my sales funnel and see where there are “holes” in it through which money flows.

The first thing I do is go to my Justclick store and see how many payments I received during the reporting period. In our case, the reporting period is one week.

Here I see the following picture:

Within a week, 39 people subscribed to the “Contactless Sales” newsletter, and I earned 8,460 rubles. Is it a lot or a little? Frankly speaking, it’s not the ultimate dream. I would like 10-20 times more. This means we need to look for weak links in the funnel and strengthen them.

Based on the justclick indicators, I see that after subscribing everything is going quite well. 4 paid invoices out of 39 subscribers is a conversion rate of 10.2% into payment. The average bill is 2115 rubles (8460 rubles divided into 4 payments). These are very decent indicators for an automatic sales funnel.

I received the most clients in a week from the “Direct Calls” source. This is exactly my blog where you are now. The conversion to purchases from people on the blog is the highest. The question is - why are there so few of them? If in a week not 29 people but 290 people came to my funnel from the blog, then my income would be 10 times more.

Maybe I have a bad subscription page and people don't want to subscribe to my newsletter? Or maybe I have too few visitors to my blog, and there’s simply nowhere to get 290 subscribers a week? Yandex-Metrica will give us answers to these questions.

Analysis of indicators in Yandex-Metrica

Following the instructions below, I set up conversion tracking on my website. These are just the first four steps in my funnel:

1. The man came
2. The person went to the subscription page
3. The person subscribed

I open the corresponding section of Yandex Metrics and see the following picture of conversions for the week.

During the reporting period (week), 3,422 people came to my site. Of these, only 56 went to the SQ subscription page (1.64%). And 23 people (41.1%) subscribed to the newsletter.

So I can see that my subscription page is working very well. 41.1% is a high figure. And yes, don’t be confused by the fact that one system shows me 26 subscribers, and the other only 23. This happens according to various reasons. The main thing is that the indicators do not differ greatly.

And here is our “hole” - a very small percentage of site visitors get to this very registration page. Only 1.64% of total visitors.

This is very little. This is where we need to work - do more persuasive appeals, make a brighter and more catchy banner, and all that jazz. And of course, we need to work on increasing site traffic. If I had 3422 visitors not per week, but per day, then this would greatly increase the final result.

Now you can see how end-to-end analytics can help you make decisions in practice. If I didn't have all these metrics, I might think that I have a bad email series and that's why people don't want to buy anything.

Or that my subscription page is bad - and I need to completely redo it, or even change it free product, which I offer for a subscription (and after that the entire concept of the project).

But analytics shows that you just need to make the banner brighter, and everything will be fine. This is what I will do right after I show you how to set up the same end-to-end analytics system yourself.

Setting up end-to-end analytics in 3 steps

As agreed, let's look at a more “traditional” version of Internet business and analytics.

Let's say you decide to create a small business in order to finally leave your hired job and become a free person. As a niche, you have chosen the sale of soft toys, sewn by skillful Chinese hands. A normal niche, no better or worse than others.

You've completed a couple of "success" trainings and know that first of all you need specific goal, to which you will go. And with this goal you set yourself an income of 300 thousand rubles per month. According to your calculations, this should be enough for the first time.

In order to earn 300 thousand net, you need to make 600 thousand turnover per month. Because half of the income will be eaten up by taxes, overhead costs and the cost of goods. And there is still the issue of advertising. Because it is completely unclear how much it will cost you.

But having estimated, we decide that 1.5 million turnover should definitely give us the desired 300 thousand rubles of net profit per month. We sell Chinese bears at a price of 2000 rubles per piece. Therefore, we need to sell 750 bears every month (25 bears per day).

These are our initial data, and now we need to set up the entire sales funnel, immediately “implanting” end-to-end analytics into it. Because later it will be much more difficult to do. And we start with the first step - choosing indicators for control.

Step #1 - selecting indicators for control

From the very beginning, we need to define control points that we will track using end-to-end analytics. It is very important to avoid cluttering reports with numbers, but at the same time maintain an understanding of the big picture.

That is, there is no need to try to track absolutely all indicators. Otherwise, you risk getting confused in them and never understanding anything.

Our funnel will look something like this:

1. A person sees an ad (Direct, Google Display Network, VKontakte Target, banners, etc.);
2. A person follows our ad to a landing page;
3. A person places an order;
4. The person pays for the order.

Accordingly, we have only 3 main indicators that we need to track using metrics:

1. CTR of ads (the ratio of the number of ad impressions to the number of clicks);
2. Number of orders placed;
3. Number of paid orders.

And in the end we get two main indicators to which all the others lead:

1. How much money did one client cost us (according to advertising);
2. How much money did one client bring us (average purchase receipt).

The ratio of these two indicators gives us the most important indicator in end-to-end analytics.

The most important indicator

This indicator is called ROI (Return On Investment). If we spent 1000 rubles to attract one client, and received the same 1000 rubles in payment from him, then our ROI is 100%. For every ruble spent, we earned one ruble and got back 100% of the money spent.

It is precisely the ROI indicator that all your end-to-end analytics will come down to. All these 5-10-100 smaller indicators should lead precisely to the calculation of ROI. So that at the end of the reporting period you could say - “So, we have ROI of 350% from advertising in Yandex-Direct, ROI of 230% from advertising in Google Adwords, and ROI of 50% from targeted advertising on VKontakte. Therefore, we are giving up the target so as not to feed these parasites, and we are transferring the entire advertising budget to Direct as much as possible.”

Now, if you can say so, then you have correctly set up end-to-end analytics, and with its help you can track which advertising source gives you profit and which takes it away.

Okay, we have chosen the indicators to control. Let me remind you that this is:

1. Conversion to order
2. Conversion to payment

And from these indicators we ultimately calculate ROI. Let's now see where and what code needs to be inserted in order to properly control the funnel.

Step #2 — Technical setup of analytics

Setting up utm tags

First of all, we need to create correct links for our future advertising sources. To track where our visitors came from, we need to insert special tags into the links. They are called UTM tags.

If you are interested, UTM stands for Urchin Tracking Module, and means “Urchin Tracking Module”. There was a company called Urchin Software that came up with these tags, and then it was absorbed by Google.

Currently utm tags are used as universal tool to track the sources of visitors to your sites. You can generate a link with utm tags, for example here.

Just come up with names for your traffic sources and enter them in the appropriate fields. I usually only use three required fields - campaign source (utm_source), traffic type (utm_medium) and source name (utm_campaign).

• yandex&utm_medium= cpc&utm_campaign= yandex-direct(traffic source - Yandex, type - pay per click, name - Yandex Direct)
• http://yoursite.ru/?utm_source= google&utm_medium= cpc&utm_campaign= kms(traffic source - Google, traffic type - pay per click, name - kms).
• and so on

This is how you get individual links to all your traffic sources. which you are going to configure. Then you can track them using Yandex Metrica (I’ll show you how), or any other end-to-end analytics system. But Yandex-Metrica is free, so we will consider it for now.

Setting up goals in Yandex-Metrica

Now we go to the office Yandex-Metrics. If your site is not yet connected to this system, be sure to connect it. To do this, you will simply need to indicate the site address and, using a simple verification, confirm that you really are its owner.

Now you need to paste this code on all the pages that your visitors will go through. The simplest option usually looks like this:

1. Your offer page (selling landing page)
2. "Thank you for placing your order" page
3. “Payment successful” page (“Thank you for your purchase”)

It is on these pages of yours that you need to insert the counter code received from Yandex. And after that, you can move on to setting goals. I prefer setting up compound goals. It seems clearer to me this way.

To do this, in the same “Settings” section, select the “Goals” item.

Then click “Add goal” and select the goal type - composite.

Now you need to define three steps that your visitors will go through. These are the three pages we talked about above.

Now you just need to save this funnel and start receiving data for analysis. After you run advertising from all sources, receive a certain number of orders and payments - you will be able to see exactly where your clients came from.

Step #3 - Tracking Funnel Metrics

Let me remind you that we specifically set utm tags in order to track where exactly the buyers came from, and not just visitors. In the end, we can see the number of clicks and visitors in the advertising system office itself.

In order to see where clients came from, we go to the section Reports - Standard Reports - Sources - UTM Tags

There you will see a list of all utm tags by which people came to you during the reporting period. It will look something like this.

As you can see, I had the most clicks on the utm tag from Yanex Direct, and also a few little things from partners.

To see which source exactly the buyers came from, select the third step of our composite goal “Purchase” in the list of goals. For me this will be the “Subscription” goal.

In the screenshot below you can see that today my most conversions came from the “fixed” source (this is a banner on my website). The conversion was 41%, as we saw above. There were also conversions from partners, but not many.

This way you can see how many visitors came to you, and from which advertising source. Next, you will enter all these indicators into your CRM or simply into excel file, and there you will see our most important indicator - ROI.

And based on this indicator you will accept further decisions about which advertising source to invest in more money, and which one you don’t need to invest in at all (or you need to greatly improve it before investing).

Conclusion

End-to-end analytics is needed in order to understand where in your sales funnel there are “holes” through which money is leaking. Too many businesses fail because they make bad decisions because they don't see the whole situation.

They invest money in advertising that doesn't work. They're redesigning landing pages that were actually converting well. Or they even close projects that would generate excellent income, even after minimal changes to the funnel.

I hope that this instruction will help you gain “X-ray vision” and make only the right decisions.

Save the article to your favorites and share with friends. Don't forget to download my book. There I show you the most fast way from zero to the first million on the Internet (extract from personal experience in 10 years =)

See you later!

Yours Dmitry Novoselov

What does every business owner want? Know exactly the amount of advertising expenses and understand how much money you managed to earn from each advertising method. Standard analytical methods (clicks, impressions, transitions) do not provide the necessary amount of information with which you can determine effective and ineffective channels for attracting customers. These problems are solved by end-to-end analytics.

Order website promotion

Ingate in the new issue
"Transformer"!

Watch the video and learn digital tips from the company’s top officials.
Find clients. Faster!

Three levels of analytics

Given the analytics frameworks described above, three levels of complexity can be defined.

First: everything by hand. A suitable option for small and micro businesses that cannot afford paid analytical solutions. It requires the following set of tools:

• Spreadsheet Google, Excel;
• electronic platform for receiving applications (CRM);
• Google Analytics, "Yandex.Metrica". You can use the services in a complex;
• call tracking system (automatic or semi-automatic).

Action scheme: the analytics system records user actions on the site. When the application is submitted, information about the source enters the CRM. In the elementary model, the tools are not connected, so the data must be uploaded manually, combining them according to the required parameters. This is convenient for low traffic and budget. As soon as the company overcomes this stage, it needs to choose a more complex solution.

Second: automation. A solution for small or medium-sized businesses with a large volume of operations. Suitable tools:

• Google Analytics;
• automatic call tracking system;
• tool for automatic import of expenses;
• if necessary - Microsoft Power BI.

The implementation period takes on average 7 days. The principle is based on setting up a connection between tools for automated data upload to Google Analytics. Most modern systems support this feature.

Despite all the advantages of automation, this solution has its own characteristics. There is no access to each user within one session. If the number of sessions exceeds 40 thousand per day, the system can sample data. In addition, the system prohibits the transfer of personal information.

Third: perfect analytics. This level is chosen by representatives of medium and large businesses. It allows you to automate processes without the problems mentioned above. Medium Tool Set:

• Google Analytics;
• automatic system call tracking;
• database (preferably cloud);
• Excel or more advanced tools (Tableau, QlikView).

The principle can be described briefly: all received analytical data automatically enters the single base in its pure form without distortion. Thanks to this, you can obtain a sample for any request: study the behavior of each specific user, analyze groups of buyers, test great amount marketing hypotheses. There are no restrictions. The obtained data can be downloaded in the form of convenient tables and graphs.

Three stages of setting up end-to-end analytics

When implementing a system, the services of programmers and integrators are required. When everything is adjusted and tested, you can use the tools without involving IT specialists. It is better to implement end-to-end analytics when building a sales funnel. Without this, it will not be possible to estimate advertising budgets. In general, all work can be divided into three stages.

01.

Selection of key control points

The number of business indicators available for tracking can be unlimited. At each stage, these indicators may change. You can select those that are relevant for a given time period. For example, key indicators during the period of attracting customers there may be clicks on advertisements, at the stage of sales growth - the average bill and conversion, and during retention - repeat orders.

Among the most important indicators you need to take into account ROI - return on investment rate. This is the amount of money that was returned by spending money on advertising. In essence, end-to-end analytics comes down to calculating how much the company earned through advertising and how much each channel pays off.

02.

Setting up analytics

Integration of used services into unified system analysts. Setting up UTM tags and generating a link that can be used to track the sources of visitors. Customize link markup to get accurate information for each channel.

03.

Metrics tracking

Recruiting a mass of events and visitors, tracking channels using UTM tags, using analytics and call tracking systems to analyze target actions. All indicators are fed into a unified analytics system, where there are tools for their convenient calculation and forecasting.

Market leaders choose Ingate

How to become an Ingate client?

To order a strategy in Moscow or another city, simply fill out an application. A customer service manager will contact you and prepare a personal Commercial offer performance-marketing agency Ingate, compiled taking into account your goals and capabilities.

Today we will talk about end-to-end analytics and how to implement this tool inside Google Analytics.

What are we going to talk about:

Why do you need end-to-end analytics?

How to implement it in practice?

What errors can there be and how to solve them?

What is end-to-end analytics?

What situation might you have - a large number of different advertising channels through which users find your site. These may be channels that we manage, such as paid advertising, contextual advertising or special communication channels. This may be advertising that we manage indirectly - for example, search engine marketing in search engine results. These can be any advertising channels.

Simply put, we invest some money, for example 1000 USD, in our marketing. These 1000 dollars are distributed in some shares into different communication channels:

$300 for search engine optimization,

$100 for price platforms.

That is, we can distribute these amounts evenly or unevenly. And we need to understand how effective our investments are, and which ones work out better. For example, where is it better to invest - in SEO, Yandex Direct or Google Adwords. Or close everything and work only with price platforms or some aggregators. Only after analysis will we understand when there will be a greater return on the money we invest.

How to calculate the return and efficiency of these investments in different advertising channels?

On this moment Most businesses or marketers who work with Google Analytics consider cost effectiveness solely based on some kind of user communication actions. Those. the user can call or leave a request, or even, if this is the case of e-commerce, go through the entire ordering procedure and at the last step click “Place an order”. It's great if we have the ability to accept payment from the user at this step or at one of the last steps. When the user has the opportunity to enter his/her data credit card or pay via PayPal. In this case, we clearly understand that these are the people who placed the order, these are the people who bought from us.

This is ideal. But in reality? The number of transactions on the Internet using cards or other payment instruments is relatively small. I have statistics for last year or the year before, something in the region of 3-5%, this is not even a dominant percentage, i.e. it's not 50%. We understand that a large number of audiences pay whatever they want.

And the fact that the user communicated (filled out an application to purchase a product or sent us a letter or used an online chat) does not tell us about the fact of the purchase. It is rather a certain desire to buy, a certain step before a real purchase. And if we look carefully, and for your real businesses we take the number of these communications (i.e. the number of times where each individual user contacted the business with the desire to buy) and the number of purchases we will see: - the number of communications we have is much greater than the number shopping.

For example, 100 users went through the entire checkout funnel and only 50 of them actually paid.

Users may not pay for various reasons - they may not like something at the last moment, there may be a million situations why the user did not buy, but, nevertheless, such a situation arises quite often.

What is displayed in Google Analytics?

In Google Analytics we display (if the goals or e-commerce block are configured correctly) information about transactions, products that the user has selected. In this case, we see, sort of, sales. But these sales amounts or the number of transactions that we have inside Google Analytics may not coincide with what we have in CRM. If so, then we may actually make incorrect decisions regarding the effectiveness of these channels.

For example, we see that we have a lot of people leaving a request or filling out a free search basket from Yandex, Google, but at the same time the number of people who paid is much less. In this case, when we calculate the effectiveness of this channel, if we only take into account those actions, only applications and not actual fees, we may be greatly mistaken in the effectiveness of this particular traffic source.

Maybe we, say another Google Adwords source, have a smaller number of completed purchase requests, but at the same time the conversion into real payment there is quite high (100 people go completely to the shopping cart with Google Adwords, 95 of them pay). And we understand that in this case this channel will be clearly more effective for us than free search in Google if we work with the tool

The second point is the moment of taking into account all communications. Users can contact us different ways. We can quite easily and simply track the fact of communication, i.e. what happens directly on our website. If the user left, filled out the form feedback or has completely completed the ordering procedure, we will definitely see that this particular person has done something useful for us.

But if the user called us, sent an email to our regular mailbox, if he came to the store, if he used online chat, then by default all this information is not collected. And for very large quantity For businesses, we may have a situation where the number of these communications through call, letter, offline and online chat will be more than 50-60-70%, and the number of people who fill out the form may not be very large, but in fact we are the only ones and we record only those people who filled out the form with us.

Let us formulate three big problems:

1. We do not see real payments from those linked to advertising or non-advertising channels that we use.

2. We do not take into account all communications by default. Those. Thus, we cannot compare the data that we have in CRM - that this particular order came from this particular traffic source and the user used, say, a call, a letter or something else. If so, it is quite difficult to make decisions about the effectiveness of advertising channels inside Google Analytics. And then how to make decisions about the effectiveness of a particular channel, advertising campaign inside Yandex Direct. and in general any entities that we transfer to Google Analytics (regions, social media, any other information) are difficult. We actually don't know any of this. In fact, the data we have inside Google Analytics cannot be directly trusted.

3. We do not take into account returns

By law, in different countries Within 14 days, the user can return the product without giving any reason if it remains in presentation. If you have a business related to clothing, if you have a business related to some things that may potentially not suit the user, then the number of returns may be large. And by default they are also not displayed unless this is implemented additionally.

1. the user did not understand how he contacted us, or rather, we understand, but we cannot connect this with the advertising channel

2. we do not see the exact payment

3. We do not accept returns

The analytics process works as follows - we have some kind of advertising, this advertising generates some kind of traffic, users who visit the site, and actually some of these users become leads and they go to the sales department.

But in fact, in order to correctly collect data and correctly take into account the effectiveness of traffic sources, we need a different scheme. We have traffic that is generated by visitors. These visitors can use a variety of tools to communicate with businesses, and We must take each of these tools into account within CRM. But at the same time, we must understand inside CRM that this user who used the online chat came from a specific advertising campaign, for example, Google Adwords for some reason key query. And as soon as a decision is made here in CRM, i.e. in fact, the manager who is handling this order will set the status that this client has paid, we immediately transfer information about this successful payment to the analytics system. We can also transmit all sorts of communication things in order to build funnels and see the effectiveness of not only advertising tools, but also other methods of communication.

What does the new analytics system look like in practice?

We may have such a situation - we have implemented a callbackhunter, which gives us a certain number of leads and applications. And we think - oh, how cool, great, good tool, we will pay, but if we see the overall picture of the redistribution of ways to contact us, we will see that this callbackhunter can simply take a bite out of other channels - phones, forms, online chat

Thus, before the implementation of callbackhunter we had 100 calls per month and after the implementation of callbackhunter the same 100 calls per month. Those. it did not bring any additional value to the business

Accordingly, from the site, by default, data should also be transferred to analytics systems, but here more information about how the user interacted with the site itself, how he looked at various sections, how he looked at the product card, how he clicked on some buttons that were important to us, and so on.

And after we take and implement such a scheme

If we take into account each communication channel, if we take into account returns, if we only transfer data from CRM to Google Analytics from the moment of actual payment by the user (only at this moment) - only then can we build such beautiful funnels. And then we will have a complete picture of the effectiveness of our marketing. We may have different advertising channels that turn into applications, i.e. users make some kind of communication, then at this stage we see the situation regarding the effectiveness of advertising channels in terms of applications. We see how much one application from Yandex Direct or one application from VKontakte or one application from somewhere else costs us. We see the number of applications, we see the conversion rate (i.e. the ratio of those people who left an application to the traffic that came to the site in general).

Actually, we can see deeper, we can see the deal. This is not just an application, but let’s say a transaction, where some kind of communication between the manager and the user has already taken place, they draw up some details, i.e. not just an application, but a transaction, and, accordingly, we can also transfer this information to Google Analytics.

We also have information about paid transactions, i.e. when the user actually bought. And then, here, we can see the effectiveness of the selling price. Efficiency in terms of ROI, efficiency in terms of profit for each of the advertising channels, campaigns and any other entity.

For example, we can clearly understand that Yandex Direct brought us 100 leads, but there were 2 sales, and the ROI, let’s say, is 103%. But Google Adwords brought us 50 leads, less than Yandex Direct, but at the same time we had not 2 sales, but 48 sales. Let's say the conversion from applications to sales is much higher with Google Adwords and then our ROI is 500%. Conclusion: despite the fact that Yandex Direct brings us more applications for us. but from the point of view of business, money, profit, Google Adwords is effective.

But without such a funnel we will never say this, if we do not set up all the nuances in the analytics, then we will be blind.

It is very important to see not only the situation associated with one sale. There are many businesses where we can have much more sales per user than one, and even more often a situation arises where our marketing may pay off not from the first sale, but from the second, the third, the fifth.

It is necessary to take into account all the sales that were made by this specific user who came through this specific advertising channel Yandex Direct, Google Adwords or something else. In this case, if we implement this scheme, we will have a complete picture of how to implement this matter.

We solve problems of end-to-end analytics

There are many different services that solve this problem automatically. Those. there is some kind of customization needed, some of the services are more focused on CRM, some are more focused on management contextual advertising, some are more focused on call tracking, some are purely end-to-end analytics systems. I gave three examples, but there are more. We can implement an end-to-end analytics system using these services, and they are more or less tailored to solve the problems that arise so that we can see the entire funnel, but these services have a lot various disadvantages, everyone has their own.

Probably the main disadvantage of all these services is that they are quite closed systems tailored for one narrow task, either high-quality tracking of calls and end-to-end analytics, or managing contextual advertising and somewhere end-to-end analytics. And the downside is that they are difficult to modify if we have new tasks. Therefore, today we are talking about how to implement this task using Google Analytics, namely, so that within our analytics we can see in full a report like this, where we have an advertising source, any - paid or free, where we have a completely information on the stage of interaction, our costs, the effectiveness of our ads, where we have information on profitability.

We can enter communication goals into such tables, and, in fact, inside Google Analytics we can measure each of the stages with metrics that best show us the effectiveness of a particular step. We can have several reports, each of the reports will show different metrics. Then we will see the big picture, in fact, see our entire funnel.

How to set up Google Analytics correctly?

First, you need to use advanced e-commerce functionality.

The advantage of expanded e-commerce is that we can take not everything that is implemented there, but only some part, namely the part about specific payment, and actually not take all the other functionality. That is, implement it in parts. But in this case, our implementation scheme will change. And at the last step, on the thank you page we will not have the same code that transmits information about the transaction to Google Analytics, but will be last step funnels. Roughly speaking, at the last step (transaction) on our website, we will simply transmit information that the user has completed this step, that he has somehow communicated with us.

Moreover, we can transmit this information using such functionality as check out options, not only for forms on the site, some ordering procedures other than check out, but also for any other way for the user to contact us. We can see that the user went to check out or called - and for us this will be equivalent. But just the last step - we will transfer information about the transaction from the CRM.

In order to implement this entire system we need three steps:

1) remember the user id from any communication channel. Those. it doesn’t matter how the user contacted us - he can call, he can write to us by email, he can go offline or do anything. We need to record its unique identifier from each channel. I will say a little more - this is a unique identifier that Google Analytics gives to each user.

2) The second thing we need to do is upload expenses for advertising tools. This is probably a less mandatory step, but it is important from the point of view of calculating various indicators such as ROI and so on. Those. Without this everything will work, but with this we will have much more data.

3) The third step - we transfer successful payments from CRM to Google Analytics only upon actual payment. Moreover, we transmit both payments (the fact of payments) and returns (if we have them).

Setup. Step one

If you, for example, use your Chrome browser to go to any site where Google Analytics is installed (in general, this is almost any site), not counting some social networks, click right click, click “view code”, then you will see additional block, which can be on the right or below.

Now it’s called application, before resources, where we have this tab and the Cookies sub-item, if we go there, we’ll see a cookie called GA and this unique number and there is clientID.

Those. this is a unique identifier that is issued to the user (each browser) by Google Analytics. If a user visits, for example, the same site a second time, then his unique identifier will be exactly the same. But if he comes to the site for the first time, this identifier will be different.

Our task:

— take this clientID. WITH using JavaScript, PHP or any technology we can get this clientID.

- we can transmit it along with some kind of communication. For example, if this is a procedure for placing orders through the website. Here we have standard fields: first name, last name, phone number, what the user wants to buy, etc., i.e. a lot of various information. We can add another field, which we will call clientID and, taking this unique identifier, transfer it to our CRM, thus expanding it by one additional field.

If the user called us, we can also take and get this identifier. clientID is a unique identifier to which we associate all the information within Google Analytics about this user’s visits. For example, the initial traffic source through which it came, etc.

If we see this clientID in our CRM, we can then transfer it and expand Google Analytics’ knowledge of what exactly this user bought. Accordingly, we have different communication channels - call, letter, online form, online chat. Each of these communication channels can be taken into account.

Call- here we need dynamic call tracking, with which we can replace the phone number for each client, user, who is recorded in Google Analytics. Accordingly, upon a user’s call, we will be able to obtain a unique clientID from the call tracking system.

Letter we can send to the address [email protected]. The letter will be sent to [email protected], but at the same time we will retain our identifier. Those. we will show each user a unique mailing address, but all mail will be sent to our main mailing address [email protected].

If this form, then everything is very simple here. If a user places an order with us through the website or submits a form, then along with his contact information we also take the clientID, a unique identifier, and transmit it.

If this offline, then we can display a banner with a QR code on the site, in which this clientID will be encoded and oblige our managers to ask for the presence of this QR code, if it is, then we read it as usual mobile phone with a camera. Thus, this code will be saved with us.

If this online chat, then here we have the opportunity to transfer some custom field to the administrative panel, which the sales manager works with - there will be a name, correspondence, phone number and, among other things, clientID.

Setup. Step two

There are many services that allow you to do this. Most of them are paid. For example, OWOX BI - they have the ability to import data on advertising costs.

Setup. Step three

The third step (the most important) is transferring data from CRM to Google Analytics. This step is implemented using Measure Protocol technology. This is one of the APIs that Google Analytics has internally. With this API we can expand the data we have. Those. data, on the one hand, is collected from us using regular code that we place on the site. And another way is to transfer data from any place where we have Internet. This could be a CRM, any website or device connected to the Internet. From there we can transfer some information using the Measure Protocol to Google Analytics. In particular, in solving our problem, we can transmit information about the successful transaction that the user made.

The last step. Analyzing ready data

In this case, we receive ready-made data - we have a source, channels, sessions, costs, profit, price per transaction, number of transactions, number of conversions and ROI (already ROAS). We can add some other necessary and important data to such reports, but, in any case, here we can fully see the situation that we have. And then we can clearly say which advertising channel is more effective or less effective for us.

We can visualize our reports. For example, we can use Google DataStudio to display all this information visually and work with this information and conveniently analyze advertising campaigns.

When transferring data via measure protocol, a new session is created within Google Analytics for this user and this session with direct/none.

Problems arise here from time to time, because the attribution model that exists inside Google Analytics is the last indirect click, and if the user has already visited once through direct, if we have a long enough sales procedure, he could come back to the site after the initial advertising channel and , accordingly, the last indirect click will not be our initial source through which he made the conversion, but, say, another channel or the same direct.

What can we do?

We can overcome this in the following way - we can, along with the clientID, transmit the value of this metric if this is a paid advertising channel, or the value of the source and channel of traffic that the user used when he made the communication.

Attribution only works for channel sources, not regions or other session variables.

If we need these additional characteristics, for example, where the user came to us from and what he bought, from the point of view of the city, then we can also transfer this information to CRM, we will link this data, and then our transaction will be recorded exactly in the region where the user was when he left the application.

Measure protocol may not always work correctly and this must be controlled. Periodically, problems arise with data transmission via this protocol if we use POST requests, often fewer problems if we use GET requests. Experience suggests that when implementing this scheme, it is advisable to have a log file in which to record each request that was sent by Google Analytics, because Google Analytics may simply respond that everything is fine, but in reality everything may not be fine. If we have a log file we can quickly find any errors.

If the CRM is a mess, then Google Analytics will not help.

For example, if we have currencies inside CRM written differently - somewhere we have rubles, somewhere dollars, somewhere hryvnia, somewhere euros. And, if we do not reduce this data to one form or do not take it into account at all when implementing the entire scheme, then Google data Our analytics will be inaccurate.

The same applies to the payment accounting process itself - if we have many different statuses that we have or may have upon successful payment, then here we also need to be very careful when implementing this entire scheme, so that there are no moments when the payment is fixed, but in fact, the money has not come and will not come.

You receive many calls and requests every day from potential clients. End-to-end analytics helps you know exactly which of them are profitable and which are not.

The article explains what it is and how to implement it. Thanks to Konstantin Chervyakov, commercial director of Ringostat, for the methods.

What is end-to-end analytics and why is it needed?

As a rule, most people associate efficiency with ROI, ROMI, CTR, etc. For e-commerce - also the e-commerce module in Google Analytics (if it is advanced, this is quite close to the topic. However, 90% of online stores use the regular one).

All these indicators are not about end-to-end analytics. Yes, they can be supportive, especially if the sales cycle is long. From them you understand in time that everything is completely bad, or vice versa. But final decisions should be made based on real data.

This is true for any type of business, except in rare cases.

Applications and calls seem to be targeted, but they don’t bring clients. Or there are fewer sales per campaign, but the check amount is higher and includes products with higher margins. There is a risk of making a wrong decision.

For your judgment and choice - 3 types of systems, depending on complexity and advancedness.

Level "Axe"

This can hardly be called an end-to-end analytics model. But since many small and micro businesses cannot afford paid tools at the initial stage, this option deserves consideration. As practice shows, not everyone knows about it.

The set of tools is simple, with minimal, sometimes zero, budget costs.

• Google Spreadsheet, Excel;
• CRM or site admin with application sources;
• Google Analytics and/or Yandex.Metrica. The first gives more opportunities, but Metrica has its own features that Google does not have - a web viewer, reports on peaks of visits;
• Automatic or semi-automatic.

Many businesses use automatic dynamic call tracking. The system shows a unique phone number for each visitor, in order to then match the call with a specific user and learn more about him: what queries he entered, what campaign worked, what pages he looked at, etc.

Everything is available up to operating system devices.

Semi-automatic call tracking is more primitive and requires routine work:

The principle is this: each user on the site sees a unique code. The manager requests it in order to manually associate the call with a specific session.

50-60% of calls can be recorded in this way, but, as a rule, after 2-3 months managers get tired of this hassle. In this case, everything depends on the human factor.

It is important that the manager does not forget or confuse anything. This is an additional burden for him: instead of selling, he enters promotional codes. The company is losing profits.

To be fair, we note that this method is inexpensive. And sometimes it's the only one affordable option- for example, for micro-businesses with a small budget.

Implementation principle

Users come to the site and perform certain actions. Everything is recorded by the analytics system. When placing orders or applications, information about sources goes to the admin panel or CRM.

The simplest model does not have a combination of these tools, so we upload the data to Excel separately.

Here is an upload from CRM - these are all closed transactions (case from Ringostat agency):

Applications from the site, in our example - registrations, indicating sources, campaigns and keywords:

And the same sample for calls:

Excel skills make the job easier, but it still takes time. The lack of automation is the main disadvantage of the model.

At the initial stage, when the traffic is small and the budget is low, this is enough, but over time you have to expand the capabilities. Especially if the company can afford more advanced technology.

Level "Automatic"

• Google Analytics - everything is tied to it;
• CRM system;
• Automatic call tracking;
• Automatic import of OWOX BI expenses;
• Microsoft Power BI (optional).

You can leave the same Spreadsheets for visualization and reporting, but Power BI has more features and has an automatic link to GA. Cost: $9-10 per user per month.

Implementation principle

Based on requests, we set up a goal in GA - “Transactions”. Information about visits is collected automatically.

99% of call tracking services transfer data to GA.

For calls, there are WebHooks - triggers that we send in real time.

Then data on closed transactions must be transferred to GA. Popular systems - retailCRM, amoCRM - allow you to upload them directly. Either you can create your own connector in 1 week, or entrust this task to a programmer. In the end you get full control over business metrics at Google.

What do you see as a result in GA? Here is a screenshot of one of the Ringostat clients:

The Cost Per Acquisition column shows how much money you spent. “Cost” - how much you “paid” for visitors. “Revenue” - how much income it brought in.

Minuses

To understand the shortcomings of this level, let’s analyze the data structure of the analytics system:

This is how Google Analytics works: a person visits a website and generates a session. Within the framework of it, he performs actions (hits): views pages, calls, sends requests, downloads materials, etc.

At first glance, everything is correct in terms of hierarchy.

The problem is that in GA you only have aggregated data.

1) There is no access to a specific user and session, and this limits the capabilities of analytics. You cannot take a visitor and see what he did on the site, how many times he visited, what actions he performed during a certain session.

2) When large volumes traffic - about 40,000 sessions per day - there is a risk of sampling (for large projects).

This is when Google takes part of a group of visitors - a sample of 5-7% - and transfers its behavior to its own mathematical algorithm for the entirety. As a result, you see distorted data.

All that can be done is to customize the report, but this is hindered by type restrictions: one parameter cannot be included in the report with another, you cannot add more than two parameters, etc.

4) Personal information cannot be transferred. The basis is Google's privacy regulations. This further complicates the challenges of end-to-end analytics.

Level "Machine gun"

Let's look at 2 options.

Option 1:

• Google Analytics;
• Automatic call tracking;
• Cloud (Google Bigquery) or own database (MySQL, Mongo);
• Excel, Spreadsheets, Power BI, Data Studio, Tableau, Qlikview.

Your own database is more difficult to implement. That's why we're looking at Bigquery. This cloud service With high speed processing.

OWOX BI can do streaming in Bigquery. This is data interception from GA in raw, non-aggregated form. Thanks to this, you can ask any questions regarding the audience, behavior, its dependence on various factors, income for example. In general, test various hypotheses as much as you like, without restrictions.

To draw any conclusions, you need to visually represent tables from Bigquery in the form of graphs, charts, etc. Eat paid services Tableau and Qlikview - they provide advanced visualization.

Option 2:

• Kissmetrics or equivalent (Mixpanel, Woopra, Amplitude);
• Automatic call tracking.

Kissmetrics replaces points 1, 4 and 5 of the previous version. This is the so-called user-based, or person-based analytics system. All the “conveniences” are here, unlike GA: the program itself collects data, including in raw form, gives access to users, sessions and allows the transfer of personal data.

Implementation principle

If you have Kissmetrics or an analogue, you upload all the information there and connect the CRM system. In the case of a database, you throw streaming from GA, etc. into it. and select a program to visualize the results. While Kissmetrics already has visual reports.

We have released a new book “Content Marketing in in social networks: How to get into your subscribers’ heads and make them fall in love with your brand.”

Subscribe

End-to-end allows you to track the entire customer journey from the first contact with advertisement before making a purchase or sale.

To understand what end-to-end analytics is, let’s look at a simple analogy. Almost all modern offline companies evaluate and reward their employees based on the specific results of their work (fulfillment of the plan, sales volume, volume of attracted clients). That is, the work is not assessed by the number of calls made, the number of meetings or the volume of documents analyzed. This is necessary so that the employer knows how much the company has earned with the help of this employee whether it is advisable to continue cooperation with him.

End-to-end web analytics has a similar mechanism; it allows you to clearly answer the questions: how much was spent on advertising; how much did you manage to earn from this?

Why do you need an end-to-end analytics system?

End-to-end analytics for the website provides data from all stages of the sales funnel. This allows you not only to choose quality traffic, but also to analyze where the client was not “pressured”.

Experts identify 3 main functions of end-to-end analytics:

1. Detailed analysis of the payback of each traffic channel. Here you can calculate ROI (return on investment) not only for the whole company, but also for a specific tool (banner, teaser or even keyword).
2. Identification of the audience who made an application, became interested, but for some reason did not complete the order. This group should be directed to additional advertising, motivate her and bring her to the desired result.
3. Disable advertising for existing clients/ setting up other advertising for existing clients. This will save your budget and not irritate your customers.

How to set up end-to-end analytics

For this you will need at least Google Analytics. It is necessary to combine them into a single whole. Setting up end-to-end analytics involves 3 simple steps:

1. After the visitor has submitted an application on the site, his Client ID (client ID or cid) is transferred to the CRM system.
2. If a lead is interested in making a purchase, when the status in CRM changes, it is transferred to Google Analytics (GA). If the purchase is completed, the transaction amount is transferred to the GA.
3. In principle, the task is to enrich the GA with data from CRM. As a result, we have on our hands effective remedy for drawing up convenient reports in various directions and sections.

To turn a visitor into a lead, you need to apply 3 various formats contact. the main task: transfer the client ID after contacting him.

Possible contact types:

• Standard application, order form.
• Telephone. It is most often used by users who prefer to place an order with an operator over the phone.
• Online consultant.

To set up an end-to-end analytics system, you will need the following tools:

• Web analytics system (GA)
• CRM system
• Integration service (for example, Albato.ru)
• Service
• Online consultant
• If necessary, additional forms of interaction with clients.

Step 1

First of all, we set up a function in CRM to transmit the cid of those visitors who have shown interest. This cid is assigned by the GA and is unique for each user; it is presented as a digital combination, for example: 758596365.0236524158. The Cid is placed and permanently stored in the Internet browser cookie; it is assigned to the visitor by those sites where the GA is installed. Now you are looking for yourself useful information and read the article, and in the meantime, the site has already assigned you a cid. To find out your cid, press the F12 key, then select the “Application” section, then “Cookies” and you will see your cid.

There are several ways to send cid to CRM. If the site has a contact form, you can embed an additional item into it (it will be hidden) to automatically send the cid to the system.

Before this, in the CRM system itself, you need to open a field where user cids will flock. This can be done by a webmaster or programmer. If it is not possible to use their services, use a ready-made CRM form designer, which already has the option to send cid. For example, amoCRM (note that here you need to check the box next to GA.

Now our order or application form is integrated with CRM and transmits cid. If the site has additional channels communications, they also need to be integrated with the system. Most call tracking services and online consultants have an option for integration with CRM in the settings; just make a couple of clicks (to find out which clicks you need to make, read the help of the application you are using).

Step 2

Setting up data export to GA from CRM

The cherished cid, which is talked about so much today, is the key to linking the history of actions and site visits. You can configure the export of any actions based on your business needs. For example, consider export next steps: change the status to “Application confirmed” and change the status to “Transaction completed” indicating the transaction amount. We will set up the connection using the unified integration service Albato.ru (small companies may well limit themselves to the free version, but for large businesses it is better to purchase a paid version with a payment of 999 rubles/month).

First you need to register with the service and add the amoCRM - GA link.

After this, a new window will open where we specify the connection parameters. In the “amoCRM” item we set “Transaction status update”, in the GA item – “Event”.

In the third step, select the connection that you created before.

• Tracking ID is the number of the counter used in GA
• Client ID. This is the same cid. Client IDs were collected in advance and linked to clients in the CRM system.
• Event category. This can be any parameter, the main thing is that it is more convenient for you.
• Event action. Any name can be entered here too.

After that we save changes made and we see our link:

amoCRM:: Deal Status Update > Google Analytics:: Event

Next, we specify in the filter the events that we will transmit. For example, you want to view in GA the source of traffic that brings confirmed transactions. To do this, change the link by clicking the pencil icon. Next, click on the icon in the middle.

A window will open in which you need to enter the necessary data, namely, we change only one line. We specify that when the status is changed to “Confirmed transaction”, the event should fire (in our example, the number 14843968 corresponds to this status). Save the change.

As you understand, we need to know the ID of all statuses of the CRM system. To do this, in the same Albato, open “Connections”, go to amoCRM and select the “Edit entry” item. Next, a window with statuses will open.

Next, “Start the connection” and look at the status. It should change from "Suspended" to "Pending". If the status has changed, everything is done correctly and the link is attached. A second link is created in a similar way so that not only order confirmation events are transmitted to the GA, but also completed sales indicating the amount.

As a result, we have 2 working links.

Today there are various end-to-end analytics services; you shouldn’t limit yourself to Albato. You can try Roistat, PrimeGate.io, CoMagic. End-to-end analytics + Comagic Call Tracking, K50 and so on. Some services are paid, some provide free versions for small companies. Try different services and choose the functionality that is more convenient for you to work with.