Methods for evaluating call tracking systems: thoughts on the topic

    One of the buzzwords of 2015 for marketers of all levels was the word “call tracking”. This technology allows you to significantly increase the number of targeted calls from advertising without increasing the advertising budget. And it is quite logical that it has become fashionable to conduct and post studies of such systems.

    In this case, most often I have come across custom articles that are easily identifiable by even a not very prepared reader. Another option is surface studies using fictitious cases as an example (a spherical site in a vacuum with N people per day and session time M minutes), for which the cost and / or several other parameters not directly related to reality were estimated.

    Unfortunately, I have not yet met a quality independent study. Though take it and do it myself, by golly.

    image


    I want to express my thoughts on what a qualitative study might look like. Who knows, maybe someone will volunteer to carry out such a large-scale project.

    “And who are the judges?”: About users


    We will conditionally divide the system users into 3 categories:

    1. Beginners.
      These are those who have heard about call tracking and understand how fashionable it is and what needs to be implemented. But their technical base is lame, and they don’t want to pull it up, because they’ve been looking for the “do everything well” button all their lives. The main thing for them is to get beautiful reports and a simple interface. Finer points are not interesting to them.
    2. Advanced users.
      They delve into the service and ask the right questions. Find out the principles of pricing and why prices are different for different market players. Understand the differences between price and value.
    3. Experts.
      This is the guru of analytics and their immediate environment. They dig very deep. And, sometimes, they know the product better than the product developers themselves. It is easy to tell all the disadvantages of each of the systems and give specific recommendations on how to fix them.

    Actually, the article (and research) for users from category 2 and above. But readers from category 1 will be interested to think about whether such a button exists in the call analytics:

    image

    General points


    Call tracking is a complex product. He has many pitfalls . And each player in this market has its own "chip", which is not always easy to see.

    Therefore, the necessary stages of a qualitative study are:

    1. Attracting a technically competent expert to the study.
      It is not only about a good understanding of the algorithms, but also about the expertise in Internet marketing. Since call tracking is, first of all, an analytics tool.
    2. Deep immersion in the algorithms and technologies of each service.
      And this implies, at a minimum, the presence of a specialist certificate in those systems where certification is available.
    3. Discussion of the research methodology with representatives of each of the studied systems (otherwise, again, the study risks not being unbiased).

    The study itself should also be carried out on the example of a real case. In order to solve specific problems, and not just write a report. In general, research for the sake of research should inspire more suspicion than credibility. It's like drinking good brandy to get drunk - a waste of resources.

    Well, what is important to research, besides the price and quality of work of sellers?

    Accuracy


    What is the cornerstone of any analytic tool? Correctly, the accuracy of the data that can be obtained with it. After all, on the basis of these data, I will make some decisions: cut budgets, look for the guilty, and reward those who are not involved.

    In my understanding, the accuracy of the call tracking system is how many calls are CORRECTLY connected with the visitor (here, of course, we are talking about dynamic call tracking, everything is clear with statics).

    The parameter that shows how many calls are generally connected with visitors is Coverage, not accuracy. Even if someone is trying to replace these concepts, do not believe it.

    To assess the accuracy should be a technique. Clear to any potential user. Moreover, the statements “we know this, because we have a lot of experience”, etc. - This is not a technique, but an attempt to hang noodles on kindly provided ears. A reference to big names for which everything works well is also an attempt to change the subject. The technique is either there and it is clear, or it is not.

    And, besides, this technique should be scalable, and more on that below.

    Scalability


    If Vasya from Razbrod i Reeling LLC receives 100% correct data with your service (which is already strange for an analytical service), for me the world may not be so beautiful.

    I want to know how accurately I will receive the data! Those. if the system has a methodology for assessing accuracy, then I should be able to apply this methodology to my site and get a forecast for it, and not the average for an industry hospital / market / solar system. After all, each of us is unique. Not to mention our sites.

    image

    Broken phone


    I remember one of my first phones was the Nokia3310. And it was the only phone in the reliability of which I was sure.

    What is it for? Call tracking, in addition to counting calls, also directly affects their number ( here was an article on this topic ). Those. if you replaced your old, but working number on the site with a new but not working call tracking number, then there will be less calls. The opposite situation is also possible: numbers from a better operator will help your potential customers call you more often.

    In a study, it would be good to observe this point more closely as well. At a minimum, keep a log of problems with telephony: date, description of the situation, timing and results of the decision. It's minimum.

    And it’s better, of course, to conduct a couple of stress tests of swing numbers. But this is a completely different story and draws on a separate study. Therefore, we restrict ourselves to the log.

    Service transparency


    All systems that have something to hide are like McDonald’s: it’s like meat at the entrance, meatball at the exit. And between these conditions there is something, because of which you can quickly get an extra ten kilograms of weight, shortness of breath and problems with rational thinking.

    Therefore, exploring a system that will tell me where to redistribute my hard-earned 50% of my wasted budget, I want:

    • Have the opportunity to learn how each of the components of the system works.
      To understand how the work algorithm meets my needs.
      Beautiful reports are, of course, good. But for that kind of money, I can draw more beautifully. So I would like to know where they come from and whether they can be trusted at all. And, ideally, he’ll pick something up himself, because only my wife knows my business better than me. And even that is only his financial side.
    • Promptly receive notifications when something breaks down in the system.
      To understand how this will affect my business.
      It is sad to find out after two months that, it turns out, once a week a third of the rooms do not work for me.
    • Be able to track how the problem is being resolved by the service provider.
      What work is being done at any given time. In order not to bite your nails in "naive ignorance", but to quickly make a decision "how to live on."

    The last point is closer to customer service. I’ll talk about it a bit later, but now I’ll try to develop a simple criterion for the value of the service. After all, the avaricious pays twice, but we do not need it.

    Price / accuracy criterion (price per unit of accuracy)


    image

    According to Pareto's law, an increase in the accuracy of a service at a level of more than 80% leads to a disproportionate increase in its price. Therefore, in addition to the accuracy index itself, it would be nice to keep in mind another simple criterion. Without hesitation, I took this: the price per unit of accuracy (ZET).

    We take the price of the service (for example, 20,000 rubles) and divide by the percentage accuracy (for example, 80%).

    Total, ZET = 20 000/80 = 250 rubles.

    If, at a cost of 15,000 rubles. we get only 50% accuracy, then ZET = 300 rubles.

    Therefore, even if the cost of the service is lower, it is premature to say that it is of better quality, at least.

    I came up with this criterion in 15 minutes. It doesn't have to be that way. It all depends on the tasks and the desire to understand this difficult world. The criterion, if you get to the point of absurdity, may be the number of phrases "we are the best" per unit time from the seller.

    Customer service


    And, of course, do not forget - a high-quality provider of a service vital for business (and the service through which my calls go is exactly that) should have several mandatory points when working with clients:

    1. More than 2 communication channels: telephone, mail, and something else (personal manager’s mobile phone, skype, social networks, etc.). So that I can reach 100% respected colleagues in the event of a big bad boom. And bad booms will certainly happen, because the more complex the system, the higher the probability of error. This is a law of nature, there is no getting away from it.
    2. Clearly defined deadlines for working with customer requests (for example, response to a mail request is up to 15 minutes, a call is instant, in ICQ - once a year). There should also be deadlines for feedback on the issues that I hand over to technical support (for example, for urgent applications - 15 minutes, for non-urgent - up to 2 hours, for proposals to drink a coffee - 24 hours). This approach removes the fear of the unknown and saves a lot of nerves.
    3. Transparency of the solution of problems (this was described above). Ideally, this is a ticket system directly in your account, where real-time comments appear in the course of solving my problems. This approach saves me the need to call and every 30 minutes to calmly explain my problem to new employees of the company.

    Actually, in short, that's all. Of course, the taste and color of all the markers are different and it is unlikely to be able to find a universal solution for all tasks. But such an approach to research will at least allow us to identify the segment for which this or that instrument will be optimal.

    Good luck with your research and only independent expertise!

    Also popular now: