Russian online furniture store Hoff keeps track of calls from the site
Literally a few years ago for the majority of companies there was a huge problem — the inability to evaluate the effectiveness of your investment in advertising. Primarily, this problem was related to the fact that almost every business has multiple entry points customers (shopping cart on the website, the application, the call). And if with online shopping all have learned to work through the module configuration of e-Commerce in Google Analytics and “Yandex.Metric”, then calls the situation was much worse.
That though as-that to evaluate the effectiveness of invested in advertising tools, website owners have introduced additional challenges for operators of call centers, in which the operator had to figure out how the person came to the site and to incorporate such information into CRM. It is clear that such an approach gave low accuracy, but it can be considered the progenitor of modern call tracking.
In order to increase the efficiency of collecting information about the sources, came up with the idea to post on the website next to the phone unique identifier which can be assigned to each user accessing our website.
Every time the client made the decision on registering with the site, the operator is asked to call the given identifier, allowing to directly compare a phone call with the user session. Further, this information was managed in CRM. The method was all good, except that implied a heavy burden on operators. As a result, even with very well-established system of making IDs in CRM, the total error in determining the sources were obtained at least 30% — because of the human factor.
Next came the understanding that it is necessary to gather information, but to absolve the operators of this area of responsibility. As a solution to the problem was proposed methodology, when a telephone number was assigned to a specific source. That is, if the site owner can put their ads in “Yandex.Direct” and Google Adwords, it could hang on a website script that would define what source man came and showed him the corresponding number assigned to the source.
This approach solved the problem of dependence from human error when delivering information manually (as in the case of IDs) and was named the classic call tracking. However, the technique has provided very little information. Agree that understanding that “Yandex.Direct” brings you 20 calls, Google Adwords — 19, not enough to optimize the cost of advertising. You need to clearly understand what queries eventually bring calls, and which — no to disable those requests that do not work.
At some point it seemed impossible to find a solution to this problem, but several years ago the approach was found. There is a hypothesis that if you allocate a unique telephone number is not on the source, and for each user, we will be able to accurately associate the call with the user session. A reasonable question arises that then we would need infinitely many rooms. However, if a unique number to select only those users that are currently present on the website online, and then release them to continue to allocate in a circle to new users, it turns out that even a small pool of four rooms with 90 percent accuracy, identify advertising sources — until search queries for most sites on the Runet. This approach is called the dynamic call tracking is now being actively used on many projects.
Now let’s move on a bit from theory to practice. The benefits of call tracking services seems to be more than obvious, but how difficult and time-consuming installation process of such solution on the website?
Several years ago, the implementation of such a system really took a huge amount of time, but the current product is installed on the website as easily as any of the counters analysts. You need to perform only two actions:
1. To install JS-code counter on the website. Below is an example of such code:
2. To mark up in the HTML code all the places where Dating telephone number, special ID. Below is an example markup:
The installation process of the service usually takes no more than 10-15 minutes. After done these two simple steps, the service begins to collect information on calls. Remarkably, regardless of the type of call tracking, which is installed on site, the installation process is fully identical.
In this theoretical part, I want to finish and seamlessly transition to practical ways to use data for call-tracking to show impressive results. The following describes the case Hoff, which uses dynamic call tracking service Calltouch.
Calls on the result
There are areas of Internet Commerce, in which a large portion of potential buyers prefer to chat about the upcoming purchase by phone. How to find out who these people are and turn them into customers, we offer you to consider the example of the case of hypermarkets furniture.
“Counting those who came to the store’s website, marketers learned a long time ago. And thanks to modern web Analytics systems we can trace the exact path of visitors from our advertising to the site. So, naturally, we wanted something more, namely — find out who our buyers first saw the product and instead put it in the cart, reached for the phone. A suitable solution suggested service call tracking Calltouch, ” says Tatiana Panina, head of Internet marketing hypermarkets furniture Hoff. — as a result, we got the answer to your question and unexpected economic effect — saw the real possibility of optimizing the advertising budget”.
The time — money
The first chance we got because of data about the activity of our potential buyers during the week. Statistics service showed that they call unevenly on different days calls more or less. It turned out that potential buyers are most active on Wednesday and Thursday. Considering the time a surge of interest, we have increased the budget for contextual advertising. In the end, the calls became more.
Secondly, the data obtained helped to save time. Namely — the hours of the call center and increase its effectiveness. It turned out that the “peak activity” is necessary to span from 10:00 to 14:00. Given these data, we changed the schedule of work of employees.
A reasonable investment
Thirdly, the main source of optimization of advertising costs on the basis of the obtained data have become the tools of Internet marketing: the context and SEO.
Contextual advertising is consistently brought to our Internet store 25% of all calls. But calltracking helped to understand that not all keywords gives the result. Excluding those that were expensive, but didn’t bring no orders, no calls, we have retained a portion of the budget that is spent for nothing.
For example, an ad for “corner sofas” we had 55 calls and 18 orders, and the use of the phrase “wardrobe” gave almost zero value. To increased calls by 20% resulted in the shutdown of inefficient queries and a budget increase in the use of effective keywords.
Given these data has changed and the strategy SEO promotion, which, incidentally, brought the online store is still 25% of customer calls, nahodimsa products company “Yandex” and Google.
The main thing — time
There was a place for experimentation. Dynamic callback — one of the latest marketing trends, a tool that is used to keep the users on the site. Often, however, offer to call back appears before the visitor is not in the right time and causes irritation.
Based on the data of the service Calltouch, we made a call back. The main difference from the competitors is that now we know the exact information about how much time it takes on average an ad for a decision about the call. Therefore, the window appears at the moment when he is ready to call, but for some reason does not. In this case, often the window appears markedly different and depends on the source from which the user came from. For example, visitors with “Yandex.Direct” call us, on average, after 10 minutes of Google search — after 7 minutes, AdWords — after 5 minutes and the recipients — after 4 minutes.
As a result, callback has helped increase the number of calls by 10% per day.
In our opinion, the main advantages of call tracking:
fixation of calls in reference to the advertising source;
calculation of conversions each source in the call;
the accuracy of a key or the search request;
pattern discovery: the time of day, day of week or season.
And the results yielded the following:
Send own cases, in which you have managed to significantly improve (or worsen) the performance of the project. An interesting experiment will appear on the pages of rubrics Growth Hacks.