Less Reactive
More Proactive
173tech implements machine learning to uncover
the 'why' behind the 'what' of customer behaviour.
overview
Bespoke ML
Solutions
There comes a time when you can no longer run your business off educated guesses and market research. When you need firm answers on not just what is happening with customers but how you should affect it.173tech build bespoke data solutions for growing businesses.
Behaviour
At Scale
Automate your data flow to unlock who your customers are, where they are on their journey, their value, segment and next actions.
Predict The Future
Focus your attention on the right customers. Predict key behaviours and intervene earlier, reduce churn and improve upselling.
Not Just What, Why
People are complex! Get a deeper analysis on what is driving customer behaviour and the actions you should take.
Unlocking Value
How do you go from raw data to business value? In this short video the 173tech team discusses our methodology to doing just this.
Expert help is only a call away. We are always happy to give advice, offer an impartial opinion and put you on the right track. Book a call with a member of our friendly team today.
DATA WE INTEGRATE
Uncover The Secrets
From Your Sources
If it produces data. We can probably integrate it. Understanding your customers requires overlapping and de-duplication of data across their journey. Here are some common sources we work with.
HOW WE BUILD IT
Not Mystic ,
Just Modelled
Our aim at 173tech is to get you started with predictive analytics in just a few months. We build scaleable and adaptive models that stand the test of time.
We start off with a few sessions aimed at understanding your business model and current infrastructure to identify the gaps where data can make a difference. We map out your customer journey, different touchpoints, data volumes and KPIs.
- Your Team: Project Stakeholders (business & technical)
- Your Time: Two lots of 45 minute sessions typically
- Key Outcomes: Information gathered feeds directly into our recommendations
From here we will start an exploratory data analysis, most often focusing on the lifetime value of your customers. This analysis aims to input directly into your business decisions and highlight critical areas for growth. It also serves as the basis for us to then leverage machine learning to enact more complex models – allowing us to automate the process of this analysis.
- Your Team: Project Stakeholders (business)
- Your Time: We will walk you through the analysis and then reconvene for feedback/further questions.
- Key Outcomes: In-depth analysis on customer behaviour.
Building a machine learning model, such as churn prediction, involves several key steps. First, we gather and clean customer data, next, we explore the data to identify patterns and select relevant features. Then, we split the dataset into training and testing sets and choose an appropriate model. From here we now need to see how the model performs so we can train and refine it.
- Your Team: Project Stakeholders (Business)
- Your Time: We will present you with the initial results of the model and compare
- Key Outcomes: Data models – not yet productionised.
Training a machine learning model, involves selecting an appropriate algorithm and feeding it structured data to learn patterns. The dataset is split into training and validation sets to prevent overfitting. During training, the model adjusts its parameters to minimise error using techniques like gradient descent. We will then refine the parameters of the model such as adjusting learning rates or tree depths, helps improve performance. Once the model achieves satisfactory results, we are ready to activate it.
- Your Team: Project Stakeholders (Business)
- Your Time: Minimal
- Key Outcomes: Data models – ready to activate!
Activating a machine learning model, involves integrating it into a real-world system for making live predictions. The trained model is loaded into a production environment, often exposed via an API or embedded within business software. Incoming customer data is pre-processed to match the model’s training format before being fed into the model for inference. The predictions are then used to trigger actions, if our model was for churn risk – this might be personalised retention offers for high-value customers. Continuous monitoring ensures accuracy, and feedback loops help refine the model over time to adapt to changing patterns.
- Your Team: Project Stakeholders (Business)
- Your Time: Feedback is really key to refining the model’s results.
- Key Outcomes: Activated data models in the tolls your team is using.
OUR APPROACH
Your Fractional Data Team
Customer Behaviour projects typically start from £10k and up, and we will always start with an Exploratory Data Analysis.
We normally start with an EDA which will take just a few weeks. All we need is your transactional data and an understanding of your customer journey.
We will only need input on the customer journey for our initial work on the EDA but having a technical point of contact makes for faster requirements gathering process and ongoing collaboration. We will need two sessions with you, one to understand your business and another to walk you through our analsysis and aswer any questions.
