Startup: tariff selection for mobile

       I want to talk about one startup - - and I hope for good criticism, because we really need her now!

    image   I was always tormented by the question, what is the most favorable tariff for me. Mobile operators offer calculators to calculate the most suitable tariff, but they are based on the opinion of the subscriber, i.e. on how I think I use mobile. And I was always led into a dreary stupor by questions like “how many percent of calls do I have in less than a minute” or “how many, on average, seconds do I call city numbers”.

       Thus, the idea has ripened to select a tariff not on guesses, but on real statistics of calls (details). It can be obtained from your mobile operator. Someone’s even free.

       Without thinking twice, I decided with a friend to make a service on the Internet, because The topic seemed interesting and useful not only to us, and it is easy to develop further. We did not find similar services on the Internet in terms of functionality. What they found did not take into account many "saving" options. And we wanted to select the most profitable options for tariff options.

       Within a month, the main functionality and core were made for the analysis of detail and calculation of tariffs. Of course, a lot has changed during the development process, but, as it seems to us, they finally came to the “usable” version.

       Some technical details:
       The project itself is written in perl based on CGI :: Application using the template toolkit. Connected through the module CGI :: Application :: Plugin :: TT. This helped split the fronts of design and programming. TT turned out to be a convenient thing, but this is the topic of a separate article. All interface prettiness was implemented on jQuery.

       Considering that almost every tariff has its own peculiarities and tricks, it was decided to abandon the economy of bytes, split the code for calculating tariffs, taking the logic of each tariff into separate modules.

       As a result, we have formed a structure from the directories “city, region” / “operator”, and in each such directory there are separate files that are responsible for calculating your tariff.
       When starting the tariff calculation, all directories are bypassed and each file is connected via require with a simple code:@z= glob("tarifs/*/*/*.pm"); for (@z) { eval { require $_;}; }(eval is needed so that everything isn’t screwed up if a syntax error suddenly creeps into some file. Now the corresponding tariff is simply not calculated.)

    image   One of the “chips” of the service is the selection of the most profitable options for tariffs. It has always been difficult for me to choose which particular phone is more advantageous to assign to my “loved ones”, and it is often not immediately clear whether there will be any kind of savings from the “love” of any phone.
       To assess the profitability of the option, we compare how much the cost of detailing will be in cases where the option is enabled. If the options are mutually exclusive, use the one where the maximum benefit.
       The biggest difficulty was calculating the amount of savings when crossing various saving options. For example, if the loyalty bonus gives savings of, say, 150 rubles per month, and one favorite number saves 200 rubles, then with the simultaneous activation of these discounts, the total savings will be slightly less than just 200 + 150 due to the fact that using the bonus calls to your favorite number will cost less money, and accordingly, the amount of savings will be less. In general, not easy.

       In the process of developing the service, the first page was more and more hung with various useful information. First, the design was very simple and in fact it is nothing but a couple of pictures and download the string was not file:
    Then the news was added, instructions on how to get details from various operators
    Well, in the end, the cost graphs of the last users and the rating of the utility of tariffs were added. At one time, they even added all kinds of SEO dash - such as jokes and a list of search queries for which they came to us, but then decided not to overdo it.

    tariff rating   A few words about calculating the rating of tariffs: for each new processed detail, each tariff is given points from 0 to -X, depending on the place in the rating of profitability of the tariff. (X is the number of tariffs calculated). These points for each tariff are summarized, and the final rating level is calculated as Σ points / number of details where this tariff participated. Pretty banal simplified system, but effective and easy to program. It turns out that if a new profitable tariff appears, then it can from the very first processed detail get into the top rating. At the moment, only Moscow tariffs are involved in the rating, but in the near future plans are to create a rating for other regions.

       One of the important topics is security.. We do not ask users for passwords from personal accounts of their mobile numbers, as We believe that this is extremely unsafe. Some are afraid to share their details, because all personal calls appear in it. Now, on our part, we do not store loaded and successfully processed details on the server. The granularity is analyzed, the results are added to a pearl bar hash, and already this data structure is saved to disk (using the standard perl module Storable). It no longer contains calls or personal information, only the information that is needed to display the report. Probably, you can still come up with a system for hiding numbers before loading details, but we don’t yet know how to make it convenient for users, and besides, such that it would be possible to correctly calculate “favorite” numbers, etc.

       Monetization- our headache and the absolute lack of experience g-) But how do we make money from this excellent service? It seems to many that the service is useful, we use it ourselves, and friends say: “Wow!” But the main point of creating the majority of projects (if in garlic) is, in the end, to make money. At the initial stage, we decided to make payment by SMS worth $ 1, because heard somewhere here from a smart uncle with a beard that the only working way to pay for Internet services is SMS. At this stage, we have already revised the strategy a bit, and are thinking of registering on the site (possibly via OpenID) with one free detail and one-time payment for the whole time (i.e. with an infinite number of using the service), maybe it will cost $ 1 too) )))

       Future plans:
    • Create a flash movie with screenshots of the service on the main page that describes the use of the service
    • Continue expanding geography
    • Add tariffs for “virtual” operators
    • Add a tariff rating system (five stars) and the ability to leave / read user comments on tariffs.
    • Organize a sale with the delivery of SIM cards, if someone decided to switch to the tariff of another operator.

       P.S. As a result, when the service was already in working condition, I uploaded my details and it turned out that, as I thought, my tariff was completely unprofitable for me)))))) I overpayed 200 rubles each month, in addition, it turned out that my MTS charges 50 rubles every month for GOOD'OK service, although it was initially listed as free in the services.

    Also popular now: