All Tools> Alteryx
10 min read
Alteryx Designer is a codeless self-service ETL tool designed to be helpful for seasoned professionals while remaining usable enough for those who are less-technically oriented. It allows users to be able to combine, prepare, and blend multiple data sources, and then apply a series of analytical tools to be able to gain a better understanding of this data. From the perspective of a direct marketing operation, Alteryx can be particularly useful for spotting patterns in target populations and to be able to filter and sort them into ways that are understandable.
Alteryx comes packaged with a series of predictive analytics modeling tools, and provides a number of useful visualization tools, which can make it flexible enough for data scientists to be able to perform sophisticated analysis in a relatively short time-frame. However, being a code-free application, it is also designed to be very user-friendly, and can be of great use for less technical users, including those who lack a deep understanding of data science or data modeling.
Workflows are easily stored as modules, which enable them to be reused with other datasets. This makes it a helpful platform for creating customizable useable dashboards for non-technical users. It allows creating simple interfaces for entering and modifying data on the fly for sales management teams.
Alteryx's features can be broken out into four categories:
Users can search any data assets, and collaborate with other users, not only to create new analytics tools, but also to be able to access models that others have created to avoid reinventing the wheel.
Users are able to prepare their data and build useful models which can be used and reused for similar datasets.
Alteryx promotes sharing of information between users. They have borrowed a number of ideas from the Open Source movement and promote full sharing of information or analytics tools developed by the community.
Alteryx has the ability to be deployed directly into production, using existing R and Python models without needing to be modified, which enables business teams to be able to run streaming analytics without making major changes to workflows created in test environments.
The interface is a simple canvas with a series of tools organized and color coded by category from a top menu. Each tool is represented by an icon or node which can be dragged and dropped onto the canvas. Workflows are automatically created when placing a new node next to an existing one; Alteryx immediately recognizes an association and connects them seamlessly.
The first step in a workflow requires loading a dataset. This is handled easily by dragging an “input data” tool onto the canvas. You are immediately prompted to select a data source; this includes everything from simple flat and csv files to various databases such as SQL Server, Oracle and more, including Hadoop files and other ODBC databases, Taradata Bulk, OleDB, ESRI, etc.
Depending on the type of file, you are prompted to choose a number of parameters. A normal next step would involve dragging a “browse” tool which enables you not only to see your actual data, but also to recognize some basic patterns. By selecting a column of the data, a graph is displayed demonstrating frequencies of values within the data. It automatically recognizes various data types, such as dates, as shown below.
As you can see, this is the result of connecting only two tools; the “data input” and the “browse” tool.
Of particular use for direct mail marketers, are custom filter tools. For instance, say you are running a campaign that is targeting only US. However, your dataset includes information about people outside of your target market. By using a filter, you can choose only those within the US. The tool also creates a simple boolean switch so you can easily toggle between those for whom the indicator is true or false. You can also create custom formulas for creating more advanced filters.
In many cases, you have several different datasets, which if combined, can provide more information than you would have in only one file. For instance, let's assume that your first spreadsheet contains data regarding the location of your customers. If you have a separate spreadsheet provides a list of visits by customers to a store, you can create a parallel workflow, and combine these using the “join” tool. If both sheets share a common id, the process is relatively easy, and this creates a common file with all of the necessary data
By applying a few other tools, one can easily output the data into a .csv , database, tableau, or other formats, for use in further analysis, or you can apply it directly to the next stage in your current workflow.
As direct marketer, anything that can increase your conversion rate is ideal. Through the use of some good data analysis, you can identify factors regarding those who have responded in the past, and using this, identify key characteristics regarding this group of people. If you have a large enough dataset, it can be possible using predictive modeling to determine which groups are most likely to respond positively to a mailer, and create a model to help you in future campaigns.
Alteryx provides a large selection of predictive modeling tools created using Hadoop and R. From the perspective a professional data scientist, this can be extremely valuable. However, as Alteryx is codeless by design with a drag and drop workflow interface, these are accessible to non-experts to gain a better understanding of your data without a highly specialized toolset.
Data investigation tools include field summaries, heat plots, and more, and has built-in features for association analysis, contingency tables, distribution analysis, and various weights and correlations.
Other built-in tools include an R-based scoring tool which enables creating weights for building customer models for your direct marketing campaigns. Also particularly helpful are various predictive grouping tools, such as Append Clusters, K-Centroids, MB Affinity, Inspect and Rules, and far more.
Alteryx provides a number of industry leading tools which can help with the visualization of data. By combining a number of different datasets, you can overlay results so that you can see how it looks in a spatial sense.
For example, if one has data stored in spatial polygons, datasets can be combined visually using a built in “spatial process” tool. Using the same “browse” tool we described before can work at any point within a workflow, and demonstrate how the data looks within a map overlay.
Alteryx provides some easy to use tools for designing reports. By dragging and dropping elements, and creating basic text, one can draw out items to be used for visualization in a dashboard.
One can customize interactive models for business users. One can create new tabs, change text and add filters to enable changing views of data on-the-fly. The example below demonstrates a couple of simple tabs, with a country filter. The second tab enables specifying the number of records that might appear in a table or chart.
One of the most valuable aspects of Alteryx its community. In many cases, the analysis that a user is attempting to perform may have already been completed by someone else. Users are encouraged to share templates, so that others do not need to repeat processes that someone else has already figured out. This can make the creation of workflows considerably easier and save a tremendous amount of time, even for more advanced users.
As a result, there are many downloadable templates which may be invaluable for direct marketers. For instance, below is a piece of a Customer Acquisition Model for channel-based marketing. It combines multiple data streams, and processes them into a standard format which can be exported as a Tableau file. (It is important to note that many of these templates have some dependencies which require downloading and installing various extra modules. However, this can be done fairly easily).
While the functionality provided by Alteryx may be useful for many areas of data-science, it feels in some ways as if it was built specifically for large-scale marketing. Either by making use of pre-existing downloadable templates, or by creating new ones, this application could prove invaluable for your existing and future marketing campaigns
Alteryx is very easy to use. Most of what is necessary works right out of the box. However, while testing some of the slightly more complex functionality, it was not immediately apparent how to duplicate certain results in some of the samples. The interface is fairly easy to understand, however modifying and managing the view became a little difficult on a laptop; items often appeared off-screen and it was difficult to zoom in and out; changing the look required going into the user settings, which was a bit clunky
For this reason it does not receive a perfect score, but these are challenges that can be overcome with time, given that the overall functionality is fairly clear.
One of the factors that separates Alteryx from other closed source ETL tools, is that the community is as good, if not better, than many open source products. The Alteryx community has created a full library of use-case models, and user-created manuals and shared tools that users have developed while using the software.
Beyond the direct-user base, Alteryx provides regular support, including a full integrated set of tutorials for becoming an expert at using the tool (and by extension, data modeling and data science), through a built in “Alteryx Academy” As a result, finding assistance for solving problems is relatively painless and a strong factor in choosing Alteryx, particularly for beginners
Use-Case Library: community.alteryx.com/t5/Alteryx-Use-Cases/tkb-p/use-cases
Alteryx Academy: community.alteryx.com/t5/Alteryx-Academy/ct-p/alteryx-academy
While getting started is extremely easy, and much works out of the box, some of the more advanced features may require understanding a little more about data science. However introductions to much of the basic functionality is easy to understand, and there are many tutorials and examples, as well as a large user community which will make ramping up not insurmountable.
Alteryx's collection of features combined with its ease of use make it an excellent choice as an ETL solution. Given that it has set of already built-in predictive models designed for market analysis, it can be extremely useful for direct mail marketers wishing to gain a clear understanding of their existing markets and to be able to identify strengths and weaknesses in current strategies.
Getting started for beginners is relatively painless; much of it works right out of the box, and Alteryx enables the creation of modular workflows that can help extract, transform, and load data. It also provides the tools necessary for building useful interactive dashboards.
While to truly make use of its functionality does require some understanding of data modeling and analysis, Alteryx provides an extremely rich set of documentation, which grows as new users contribute templates. As a result, Alteryx is highly recommended for the purposes of direct marketers.
10 min read
A review of KNIME Analytics software for use in direct marketing business intelligence operations.
10 min read
A review of Talend Open Studio ETL software for use in direct marketing business intelligence operations.
10 min read
Review of Pentaho Data Integration and Analytics for use in direct marketing organizations.