Programa Comunicación Digital

Redes Sociales y la empresa

Las Redes Sociales de Internet (Facebook, Twitter, Google+, etc.) vienen cobrando cada vez más protagonismo. Hasta en el mundo de los negocios, hoy forman parte de la Comunicación Digital de cualquier empresa. En Continente Siete vemos el valor de estos canales, y hemos estado tratando de apalancarlos desde un principio.

Mantener la disciplina

Los Canales Digitales admiten altos niveles de información. Es por eso que muchas empresas optaron por crear equipos dedicados para la creación de contenido y la comunicación en los mismos. En Continente Siete buscamos la participación de todo C7er (y el tamaño de la empresa no justifica tener a una persona dedicada únicamente a estos canales) por eso intentamos un enfoque diferente.

Hace aproximadamente un año, un equipo de 3 personas dedicó sus días D (día dedicado al desarrollo de proyectos de interés personal) en entender y establecer los usos correctos de cada canal. El output de este proyecto fue el armado de una política de comunicación descentralizada, con pautas y guías para un efectivo uso, en donde cada C7er puede efectuar una comunicación oficial de Continente Siete por estas vías.

Resultados: el nivel de comunicación aumentó significativamente. Sin embargo, el costo de generación de contenido era alto y el incentivo bajo, llevando a un flujo esporádico de información que subutilizaba los canales.

Apalancando en múltiples puntos

Hace unos meses el problema recobró importancia, y un equipo de Continente Siete decidió montarse sobre la solución anterior. Esta vez, no solamente se trabajaría en la política, sino también en la organización y el incentivo.

Organización

Se usó el Trello (www.trello.com es una aplicación web que sirve como herramienta para la gestión de proyectos AGILE) y se creó un tablero para organizar el flujo de comunicación digital. En el tablero hay 3 listas: Repositorio, Programados y Posteados Recientemente. 

En el Repositorio cada C7er puede crear una tarjeta con un link a un artículo interesante, o agregar contenido propio. En esta misma tarjeta la persona se auto-asigna y sugiere a través de un código de colores los canales en los cuales cree que vale la pena comunicar.

Un encargado luego pasa estas tarjetas a Programados, en donde se le agrega una fecha de “posteo” que la persona asignada deberá respetar.

El encargado es un C7er diferente cada mes, y un Google Spreadhseet se encarga de comunicar a quién le corresponde el cargo.

Incentivo

“Gamification is the use of game design techniques, game thinking and game mechanics to enhance non-game contexts” – Wikipedia

A través de este concepto de Gamification es que se intentó agregarle incentivo al sistema. Se creó un monitor en el cual el C7er que ejecuta la comunicación debe registrar sus “posteos”. El mismo monitor muestra las últimas tendencias de los “posts” en Continente Siete, y, para cada persona, le asigna un nivel por cada canal de comunicación (por ejemplo, un C7er puede ser Level 1 Facebook, y al mismo tiempo Level 4 Blog C7). También cada C7er tiene un Puntaje y Nivel total de acuerdo a qué tanto avance tenga en los diferentes canales.

De esta manera se generan Rankings que permiten a los C7ers entrar en una sana competencia para ser el mejor comunicador.

En un futuro cercano también se agregarán Awards y premios reales para continuar en la curva ascendente de Incentivos.

Resultados y últimas palabras

Desde su implementación hemos tenido un flujo constante de comunicación a través de los diferentes canales, sin perder la calidad de la misma. La exposición ha aumentado significativamente. Si bien se agrega un costo extra que es el de registrar el “posteo”, el beneficio por facilitar la creación de contenido y por el incentivo, parece ser mayor.

Queda por re-analizar los parámetros de uso correctos de cada canal, ya que no queremos saturarlos y perder audiencia por esto.

Gracias Jose, Mechi, Migue y Juancho por ayudar a armar este programa!

Ahora, aprovecho y me sumo un puntito en el Blog C7.

Optimizing Milk Runs with WSA

Imagine your job is to schedule the pick-up and delivery of certain goods. Everyday, you will receive orders that vary in many ways: there are many suppliers from which to pick-up; different order sizes; different lead times; different delivery dates and time windows! With this information, you must decide how to satisfy the demand, using your limited resources….

It’s not a simple job! Unless you have an unlimited fleet of trucks that don’t use fuel, and a group of drivers willing to make ridiculous trips at no cost, you will be motivated to design optimal routes. Added to these obvious economic reasons, you should also be considering pollution and traffic congestion!!

OK, don’t panic... There are several ways of tackling this problem, and one of them is the Wright’s Saving Algorithm, a deterministic method that groups clients in order to reduce the total distance travelled by trucks.

The procedure is very simple, and results are suprising! It consists of pairing all orders, and calculating the kms saved by grouping these two suppliers. This image shows how in this case, grouping suppliers 1 and 2 would generate a saving: S=d13+d02-d12. 

The negative term in the saving comes because pairing suppliers adds a new route: the one connecting both suppliers. If this distance is greater than the positive terms, then pairing these suppliers will have a "negative saving", indicating that its best to visit both separately (if we consider distance as the only cost indicator).

The algorithm then analyzes all combinations, ordered by greater saving, and assigns orders to the available trucks, considering capacity and fuel restrictions.

You can try the model we developed here: http://www.runthemodel.com/models/825/

In the example, you have to locate the suppliers in the green area (by clicking and dragging) and click the START button. You will see how new orders that arrive each day are assigned to the different trucks. Try pausing the model when orders arrive and thinking how you would assign the demand to the available trucks. Then compare your ideas with the algorithm! With few suppliers you will probably be able to match the results (or even beat the algorithm!), but try adding suppliers, and the problem will become exponentially more complex!!

Milk Runs tour re-make

	Developed with simulation software AnyLogic

OF VEGETABLES AND POLISH TRIANGLES

     What do a fern and a snowflake have in common? If I were to tell you that they relate in the same way as DNA relates to Romanesco broccoli, would it strike you as odd? They are all cases of natural-given self-similarity. In mathematics, a self-similar object is exactly or approximately similar to a part of itself (this contradicts the popular saying: beautiful from afar, but far from beautiful).

     A couple of weeks ago I came across a creation of the Polish mathematician Warclaw Sierpinski and was amazed by its simplicity and beauty. The Sierpinski Triangle is a self-similar fractal figure created by following a very simple algorithm. In fact, the algorithm is so simple that you can draw it on a sheet of paper while reading this post:

1. First draw an equilateral triangle the size of the palm of your hand
2. Choose a random point inside the triangle and mark that point with a pen or pencil.
3. Randomly choose a corner of the triangle.
4. Mark a new point halfway from the last point you drew and the previously chosen corner.
5. Randomly choose a new corner (it may be the same one you chose before).
6. Repeat steps 4 and 5 indefinitely (do this a couple of times).

Note: You probably won´t notice anything extraordinary by doing 10 or even 50 iterations.

     I would recommend you draw a couple of points in order to understand the algorithm but leave the ‘indefinitely’ part to a machine. At Continente Siete we modelled this simple algorithm and will leave it for you to discover what emerges from the seemingly trivial steps you have followed.

     You can run the model by following these three steps:

• After starting the model, click inside the displayed triangle to select an initial point
• Repeatedly click on the ‘Add point(s)’ button located on the upper right of the window (select the ‘Auto Add Point(s)’ checkbox to do this automatically).
• Observe the emerging pattern (note: speed up the process by adding several points  at a time with the ‘Add points’ slider located on the upper right of the window).

     We welcome you to discover and experiment with the other features included in the model.

     Maybe the next time you eat broccoli you´ll take a closer look and appreciate the mathematics behind nature.

Winter 2012: Slalom!

Temperature gets colder, people start using (once again) their heavy clothes, sky turns grey, raindrops start falling... Winter is here! And together with it, mountains get white, and ski resorts start working, preparing the slopes for future competitions.

Our offices are located far away from those mountains, but that did not stop us from creating our own Slalom model!

In this new game, you have several skiers who compete in a first race, with three principal charasteristics:

     - velocity: it determines the velocity that each skier has to descend the slope.

     - reaction: it represents the ability to change the direction.

     - precision: a better precision gives the skier a better accuracy on aiming the following chek point.

After that first race, the best skiers remain and compete in the next stage, and new skiers are introduced, with their parameters defined as a combination of the previous winners. This way, you get an 'evolution' in the parameters, and after several stages you get a good combination of parameters, possibly a local optimum given the initial combination (of the first skiers).
In fact, if you keep running the model, you'll find that the best combination isn´t a trivial one, but instead it depends on how the circuit is made, and how the skier responds to it.

One of the most attractive features of the model is the visualization of the skiers. The default visualization is from above, but pressing the green button you can see how the skis turn on each curve and, pressing the next green button, you might even see the skier from the front.

the link to run the model is the following one:

http://www.runthemodel.com/models/812/

We hope you enjoy it!