Section 1 FastStats System Overview
Apteco’s FastStats https://www.apteco.com/ is a suite of marketing data analysis and campaign automation software. It is in use in UK, Europe, North America & Australia and has over 400+ systems and 3500 end users in all the major verticals. The platform has developed since 1987, so is a mature product. Apteco are a fantastically responsive vendor to user requests, support and understanding what end users need. Apteco produce quarterly software releases with new features, and hold an annual user conference usually based in London.
Adroit www.adroitinsight.com is a UK based reseller, with many years of experience with the Apteco technology. The FastStats system provides rich functionality for accessing, querying and visualising data and providing state of the art omni channel campaign automation. It is in fact an entire marketing platform that can provide enterprise level rollout for access and querying of a common dataset.
The system is modular with additional licensable components that suit different users and teams. Data is loaded into the system using a FastStats Designer component. This has a whole set of features for loading, scheduling and publishing data into a live system that users can access. Designer can link to multiple system sources and will support intra-day, and “real-time” interactions with data. The other major components are listed below:
• FastStats Discoverer for the analysis/selection of marketing data
• FastStats Excelsior for reporting development & publishing data/reports to end users.
• FastStats Peoplestage support complex multi-channel campaigns and supersedes Cascade, a hierarchical deduplication and campaign application, which is still supported.
• FastStats Modelling for market penetration, customer profiling, patented Predictive Weight of Evidence (PWE) scoring, decision trees (including CHAID) and clustering to help inform your targeting decisions
• FastStats Geo for understanding the geographical distribution of your Customers to increase insight, highlight areas of untapped potential and facilitate campaign planning and analysis
We review these in more depth in Section 2.
FastStats currently integrates with approximately 28 different ESPs (Email Service Providers), and a number of web analytics/retargeting technologies including Cognesia, as well as having a number of CRM integrations with Salesforce.Com and MS Dynamics. FastStats also has its own API.
The system has a multi-tier, web service architecture using a Windows.NET or web-browser based Java applet for the client interface, and is powered by the FastStats.NET Server which provides specialist marketing data analysis query capabilities. All data resides securely on the host system, and is only accessible to authenticated users. Different users can have different access models, and functionality. This is controlled via a system administrator.
The system is used via a Windows-like interface supporting pull down menus, a tool bar with tool tips, point and click selection, Explorer style trees, grids and other rich user interface elements. The user interface includes several selection mechanisms ranging from simple inclusive selections to highly powerful building block combination selections. All selections can be presented on screen as a set of data records in a spread-sheet type format, or as cross tabulations of two or more variables.
FastStats CloudStage is Apteco’s cloud hosted deployment of the system, and at the time of writing this is in beta testing and due for release later in the year.
Adroit’s View of the Platform
We are huge advocates of the platform, as it is a delight to use. The GUI is state-of the art and Apteco employ specific UX designers to improve and aid the use of the system, engaging end users as part of the programme. So you will fall in love with it.
One of the main advantages of adopting a FastStats approach lies in the combined power of the functionality across different modules, and the breadth & speed of selection, visualisation to end campaign execution. Essentially you can execute data exploration, advance analytical tasks and a level of reporting within one platform, and use audience selections in the formulation and delivery of campaigns in PeopleStage. This provides efficiencies in data handling and promotes consistent, shared and repeatable working practices across the various insight and selection teams.
With a name like FastStats, it has to be fast. Depending on underlying hardware configurations it is capable of processing and handling string searches on billion row systems in seconds, which pushes it up into Hadoop territory. However it is the type and complexity of queries it can still deliver in moments which is impressive. So we see major advantages for clients in the fast “train of thought” querying, its ability to give users easy selection, analysis and visualisation of data.
A typical issue for a Data Team will be the time spent using MS SQL to manipulate and engineer data for analytics/campaign selections using complex coding via multiple steps. In some organisations where selections can take many days or even weeks to perform in SQL; these can be carried out in one or two days. This reduction in time is a considerable plus point as it allows the data team to concentrate on other areas, rather than concentrating just on selections. A FastStats analysis and campaign solution will provide direct access to key information and will improve the speed, efficiency and automation of analysis and campaigns. The repeatability of query logic, ease of creation of VVs (Virtual Variables) and the use of wizards within the application all support this speed of creation/access to insight. Data Grids allow for easy step creation of what could be quite complex queries or cubes (if you were using OLAP type technologies). And you can build complex nested queries from other queries, or other calculated fields. The expression engine” is extremely powerful – we have implemented complex regression scores using log functions within the expression engine. The biggest issue, in terms of selections, is that most people come from a SQL background and you need a different way of thinking when using FastStats; once this new way of thinking has been learnt it is not infrequent that we hear – “to do that in SQL would have taken days!”.
This is all functionality above is within the core Discoverer Base. The other modules provide deeper or more specific functionality for analysis or campaign delivery.
So what does it not do?
It is important to remember that the FastStats platform is not a data management tool. All the data must be cleansed, consolidated and presented to the platform. So these functions need to happen outside the environment. Typically Faststats solution might pull data from a SQL environment through created SQL Views.
FastStats does not present the user with a relational model but a hierarchical one, typically with a design that places the “customer” at the top. In this sense it gets its simplicity of use and speed. Whilst a flattened model can impose restraints on querying, designs can be made which cater for Business data, and other compound entities. It is worth noting that for B2B data, FastStats has developed a "Linkages" module which for instance works with Dun & Bradstreets DUNs numbering or any pre-built hierarchical key, so that drill downs from HQ to site level data is possible in Discoverer.
In terms of underlying functionality of specific modules you will find more statistical functionality in SAS or IBM SPSS, but not the superfast querying, fantastic visualisation or flexibility. In the Modelling module you can do clustering, CHAID, and scoring, using their proprietary PWE model. You can interface the application with R for instance to export the score build a logistic regression and then reimport or use the expression engine to build a dynamic logistic model that can be updated each time FastStats is refreshed. The expression engine is incredibly powerful, but unfortunately also tends to be one of the worst documented areas of FastStats. For this reason if you are going to be using it to its full advantage you should invest in a good FastStats partner. We have built a number of models in SPSS and used the expression engine to implement them within the core FastStats database. But quite honestly if you are an out and out statistical analyst you will want a greater range of stats measures than it provides, and may need to build functions outside and then import if you want to combine with the rest of the dataset.
On the campaign side, PeopleStage is maturing as a product, and allows sophisticated rules to be programmed onto campaigns, which can trigger further communications through linkage to end channel points e.g. email, SMS, Facebook, etc. We would note that Apteco are planning to add further social channel integrations into the mix on their development roadmap.
How does it compare to other applications?
Interestingly there are a wide range of possible competitor applications, albeit with different emphases, positioning and pricing. As we might expect with data analytics and campaign software many features overlap with other applications.
At the campaign level, we are starting to see ESPs providing sophisticated programme builders even enabling other channels. PeopleStage though is still a long way beyond this with its functionality and GUI.
In terms of other BI technology, e.g. Tableau or QlikView, FastStats isn’t quite a direct comparator, essentially as it does so much more that these tools. So it does provide reporting and dashboards, but wins on it capability to do interactive powerful analysis and visualisations coupled with broad predictive modelling. Along with this it provides robust campaign management in an easy to use set of applications that can be delivered on-premise or through “Software As A Service” hosted environments. So it is difficult to compare just one area.
So what are Apteco like to work with?
In itself it might appear an odd question. We are not aware of other software vendors who have such a close relationship with their vendors and end users, and are so reactive to suggestions and evolving their product to meet demand. This now includes user design testing, reshaping how their software is presented. No software product is perfect or will suit all users, but Apteco are constantly striving for this perfection – hence 4 releases a year. They provide excellent support for the software.
Apteco uses a Partner business model in that it is the Partner that deals with the day to day questions from users; but that does not mean that they are unwilling to get their hands dirty if required. Not only is FastStats a fantastic product, so is the company. It is difficult to review FastStats as it is always evolving and improving and Apteco bring out a new version each quarter.
Many software applications try and lock the user into them and try and do everything (often badly). One of the great aspects of FastStats is that it doesn’t lock you in and force you to use its functionality. On almost every set of output, there is an export to Excel/Word/PowerPoint button and this will take the information in the current set of analysis or table and drop it directly into the respective program.
Section 2 FastStats Modules
We have provided an overview of FastStats main modules, with our views below.
FastStats Discoverer is the data visualisation tool for train of thought analysis and marketing insight. It provides powerful marketing analysis with an easy to use Windows interface, enabling analysts to gain valuable insight to improve both the efficiency and cost effectiveness of their analytical activities.
Using FastStats Discoverer you can:
• Target markets with easy data selections and cross tabulations
• Produce stunning Charts, Venn diagrams and Word Clouds
• Create customer and marketplace segments directly from visualisations
• Produce PDF reports and transfer analysis into Excel, Word and PowerPoint
• View and sort sample data on grids
• Export data in a wide variety of formats or directly to an email broadcaster
• Perform basket and transaction analysis
• Use expressions and wizards to derive new variables
• Create and save templates for quicker analysis
• Schedule standard tasks to run automatically
One historic weakness was the graphics engine for creating graphs and charts - indeed to the point that many users exported the data into Excel for charting there. This is no longer the case. The graphics engine has been updated and is a massive improvement and real pleasure to use. Visualisations and a forthcoming storyboarding feature are a real emphasis for further releases, and a valuable addition to the interactive toolkit found in FastStats. Graphs and charts are dynamic to the underlying data. So dragging off an intersect in a Venn will create an underlying usable selection. This is incredibly powerful. It is possible to build 3, 4 and 5 dimension Venns, but a 5 dimension Venn can tends to obscure rather than highlight relationships; whereas a 3 dimensional Venn is a fantastic way to demonstrate relationships that even the most rabid numerophobe could understand.
FastStats can display information on maps, but its usefulness has been reduced recently due to the demise of MapPoint, the standard application that was used to display maps, it can interface with Bing maps but until a successor to MapPoint has been found this will remain a limitation. FastStats can create drivetime and crowfly distances but probably needs to implement a batch process to improve the usefulness of this functionality.
The FastStats Modelling is the data modelling module, providing the user with the capability to predict the behaviour of customers and prospects. We think it can be a bit of a hidden jewel, and well worth taking as an optional module.
When a customer profile or cluster analysis has been created, a model can be applied which scores, ranks and segments every record in your FastStats database. FastStats offers three main modelling techniques:
1. Profiling – using a patented Predictive Weight of Evidence (PWE) method that combines widely recognised Information Theory and Bayesian Probability, this technique scores individual customers and prospects and is fast, automatic and requires a minimum of user input.
2. Decision Tree Models (including CHAID) – This method produces a set of rules which are ranked to identify distinct segments or groups which contain proportionally more of your best customers and prospects. Decision Trees are particularly good for applying to external databases. The predictive modelling tools can produce some very powerful models, either using the FastStats PWE (Predicted Weight of Evidence) or tree modelling. The tree modelling can either be carried out using a standard CHAID or a PWE model for splitting purposes.
3. Clustering – Cluster analysis identifies groups of customers and prospects with similar characteristics. This method uses the K-Means technique to allocate each record to the nearest cluster centre, enabling you to better visualise and segment your database. The cluster does give an excellent view of what factors go in each cluster, but statisticians may require more control.
By combining the above modelling techniques with the expression engine and the user has access to an incredibly powerful analysis and predictive engine.
FastStats has a very powerful expression engine, which can be used to implement sophisticated models. An example below where we implemented a propensity model we had initially developed in SPSS. The power of FastStats is that this entire underlying code can be captured behind a single virtual variable, and made available across the platform to other users.
FastStats Geo integrates seamlessly with the FastStats Discoverer interface where users can simply drag and drop the Map tool onto a FastStats workbook to create thematically shaded and point maps, or overlay a map with pie charts or histograms. You can view the concentration of your customers across a geographical region or pinpoint their exact location down to street level. Zooming in allows you to view the details of each individual record or area and these details can show any fields or statistics available in your FastStats database. There are numerous display options that allow you to present the results in meaningful ways and highlight information to the audience. FastStats Geo integrates with Microsoft Mappoint (no longer supported), Bing mapping and Open Street mapping. At time of update (July 2016) we understand Apteco have been in discussion with Google Maps as well.
Using FastStats Geo you can:
FastStats Excelsior provides a reporting environment for analysts combining the analytical power of FastStats with the flexibility of Microsoft Excel to manipulate and present data driven results. FastStats results can be combined with external data to produce integrated reports customised to suit an organisation’s unique reporting requirements. Non-technical users can interact with the report to determine which results are displayed and refresh it with a single click for on-demand reporting.
FastStats tables, queries and objects can be embedded into Excel to create a set of dashboards. These dashboards/reports can be refreshed as a whole or in parts with the simple click of a button. The reports can then be distributed around the organisation as required. If required the report update can be scheduled so that a new dashboard using the most up-to-date data is available when you come in on Monday morning.
The following FastStats analysis results can be used in FastStats Excelsior:
FastStats Excelsior features include:
2016 Roadmap Update (Added in December 2016)
Apteco produced 4 updates per year and so the application moves on quickly. Below I have reviewed some of the 2016 additions, and look forward to the 2017 features that are due to be applied.
PeopleStage, the campaign management & automation engine, has had a number of developments in 2016 including integrations with Sparkpost, Twitter, and Trust Pilot.
CloudStage was launched, and allows fast deployments of the system, and resource management through Amazon Web Services.
The following features and new application are coming in 2017:
In Discoverer the intention in forthcoming releases.
In PeopleStage the roadmap is as follows: