We developed a new class of attribution models, focusing on the real interaction between sessions in the customer journey as well as on the on-site engagement – called the Interaction-Type model or IHC attribution model.
The attribution system is complemented with an interactive dashboard, weekly analysis reports and direct API connections to the core system. The whole system provides insights into the real value generators down to lower levels such as referral domains, individual keywords and ads, etc. Product Website: Advanced-Attribution.com (in development)
For a real optimization of the omni-channel customer communication, we have to go beyond the usual attribution approach and focus the analysis much more on the larger non-converting traffic share. We have build a predictive analytics framework, based on the insights from our interaction type attribution model. The focus is on estimating the current and predicting the future communication/ interaction level of the individual user in his/her sales process. With these insights we can identify the best marketing actions/ touch points to stimulate customer transfers along sales pipeline. This is further used to maximize sales and profit.
Online media campaigns can usually very well be measured at the individual customer level, but what about the influence of offline campaigns such as print, radio or TV? We have developed a variety of models to estimate the influence of offline media on the customer journey and individual behaviors. The goal is to provide our clients with results like: “offline campaigns A and F during period xyz will increase the traffic by 10-15% transactions by 9-12%.”
Further have we extended our attribution model to incorporate offline actions as probabilistic events to the individual customer journey.
We are specialized at building reliable prediction tools for very complex cases.
Here an example case for price-curve forecasting:
We have developed a scalable and self-learning forecasting system for dynamic changing price curves. The system is developed for the context of airline tickets or hotel prices, but the algorithms can basically be applied to any setting where products are sold in advance and prices are frequently adjusted, for example car rental, ferries, etc. Case Study (PDF – English)
We’ve developed a cutting-edge SEA/SEM bid optimization tool. The tool applies portfolio optimization techniques similar to financial applications, but the foundation is an extensive bid-performance estimation engine with learning capabilities.
We also developed tools to estimate position dependent performances of search-engine ads and keywords such as impressions, CTRs, effective CPC and conversion rates. These insights can straight forward be used to SEA control and bidding strategies on daily or hourly level.
We are developing tailored data mining and machine learning algorithms to unveil the information hidden in the data. For marketing and sales the goal is usually to analyze customer data in order to predict individual interests and success probabilities of personalized offers or campaign responses. In this context we also worked on social media data with the target to predict additional information and potential response behavior to marketing actions.
Another application case Individual Ad-Placement Recommendations:
Task was to find the best advertisement to show to the right customer at the right time. We are developing custom recommendation system for advertising placements for publishers and ad-networks. Our advanced algorithms work on huge datasets and are designed for very fast computation times. The usual goal is to increase the CTR (click-through-rate) or eCPM (effective cost per 1000 impression) of display campaigns and specific creatives.
An application for the optimal workforce planning in a multi-skill service center (agents have to handle in- and outbound calls and emails). The tool combines advanced long-term traffic forecasting and computation of the cost efficient agent staffing with respect to service level targets.
Apart from project specific data analysis work, we offer our clients Statistical Data Analysis as a service:
– constant analysis of most recent data
– automated report delivery
– report adjustments or analysis focus on current needs
– regular meetings (local, video or by call)
We are also performing 360° data analysis of your business, combining all kind of different data sources and perspectives to obtain an overall picture of the business. For example combining different areas such as supply, marketing and sales to identify market trends & potentials, internal trends and problems, and to evaluate areas and processes with their influences and performance in the overall picture.
Easy-to-use decision support tools tailored to customer needs, according to the system infrastructure and problem requirements. E.g. the end-user interface is a website or also quite often Excel and the back end utilizes more advanced tools such as Python based applications. Some examples are:
– demand and price forecasting
– automated market and demand behavior reporting
– performance reporting of call center agents
– computation of optimal (balance of costs and information) updating points in integrated systems
GB processed daily by our engines.
of our clients who started with us are still our clients.
days to get us started on your project.
dedication to your needs.
Our general areas of expertise
Statistical Data Analysis
Information is the key to success! What is hidden in all the data you collect? Some ideas on what can be obtained:
customer behavior and profiles
market behavior and changes
impact of weather, seasonality, holidays, special events, promotions …
risk and profit potentials behind processes or decisions
With statistical background information, we can apply advanced simulation methods to evaluate actions as well as find optimal solution to decision problems.
Forecasting / Prediction
What can you expect from the future? – is the most important question in a decision making. The optimal decision needs to be based on the most accurate information about the future. Some of our forecasting projects:
demand (hotel, airline and car rental)
prices and their future behavior
impressions, clicks and conversion rates (online advertisements)
No forecast is 100% accurate, but modern forecasting techniques are able to provide a clear view on what you can expect to happen, in form of scenarios and confidence intervals.
Would you like to test our forecasting methods on your data? Contact us!
What patterns and clusters are hidden in your data?
We are applying machine learning algorithms, which are able to work on diverse and unstructured datasets combined from different sources, to obtain information such as:
identification of actual customer clusters
campaign response analysis and optimization (e.g. in direct marketing such as emails)
customer classification (e.g. prediction of individual customer interests and choice behavior)
In this context we apply Big Data techniques for efficient computation on large datasets.
In IT terms the question is not any more “What can we do?”, it is now “What do we want?”
With the help of cloud computing, IT resources are no boundary any more. No matter how large your data volume is, or high the data velocity or the variery of different data sources and types. – We can handle it and scale the applications to your needs.
on demand IT architectures
analytics solutions as webservices
We are mainly developing our cloud solutions on the Amazon Web Services (AWS) platform.
What are the optimal decisions or actions?
Businesses always need to make decisions on control actions or the whole process structure! Some of our projects:
Dynamic fleet routing in a network
Online advertisement (spending budgets and keyword bids)
Reservation acceptance policies
We are specialized on techniques which consider the uncertain nature of your business environment, using techniques like: stochastic optimization, robust optimization or simulation.
Pricing / Revenue Management
The goal is to maximize your revenue by optimal price & sales strategies. Key factors are price sensitivity and the right customer segmentation. It is essential to obtain a clear understanding of the underlying demand in terms of numbers of customers to expect, but also the preferences and choice behaviors of different customer types.
In this context we offer:
Evaluation of the RM & Dynamic Pricing potential for your company
Development of a tailored RM model and process structure
Evaluation of current RM practice
Boost your bottom line by changing the focus from minimizing cost to maximizing revenue!
Selection of our tools
We develop & apply advanced mathematical techniques and software tools for solving complex problems to improving your business processes.
Haensel AMS was founded in 2009 by Alwin Haensel and is a Berlin, Germany, based consultancy and software development company providing its international clients with cutting-edge analytics & big-data solutions.
We are always interested in:
highly motivated self-starters with solid analytics backgrounds
solution-oriented, real world, “down & dirty” problem solvers
“Haensel AMS is our first choice for outstanding analytics tools and services. We are especially using their avant-garde attribution system in order to optimize the real effectiveness of our omni channel customer interaction. What describes best a co-operation with Haensel AMS is ambition and eagerness for success in providing exceptional results with a clear focus on fast practical usage.”
Karel Vos / Managing Director/ Bookit
“Haensel AMS is the perfect partner for developing custom Analytics & BI innovations. When you got them on a project, they dive to the ground of the problem setting, develop some smart solution and won’t stop until they fully proofed the performance gains. The results are all the way targeted for fast applicability.”
Dogus Yildirim / Head of Business Development / MedyaNet
“Haensel AMS convinced us with their deep technological understanding and highly sophisticated solutions. After an initial setup call of 30 minutes they provided us with a tracking script that seamlessly interacted with our solutions and provided exact data right from the start. We never had a technical integration as easy and painless as this one.”