Model of a Churn

Model of a Churn

We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

Butter-making - home churns and utensils

Home butter-making took time and energy, but only needed simple equipment. Low-tech methods were still well-known in rural parts of developed countries like the USA in the mid-20th century. In the UK it became less common for ordinary families to make their own butter in the course of the 19th century, but the old ways were still used on small farms and in the dairies belonging to grand houses.

After the cow(s) were milked, the milk was left to settle in a cool place, in shallow dishes, also called setting dishes or pancheons, so the cream would rise to the top. (Unless the butter was to be made from whole milk: less common than making it from cream.) Brass and earthenware dishes were used in the UK in the 17th and 18th centuries, with earthenware becoming gradually more popular, as brass sometimes tainted the flavour.

After half a day or so, the cream was skimmed off and put ready for the churn. Small home producers would want to collect a few days of milking to have enough cream to be worth churning, and a little fermentation would "ripen" the flavour. But the cream couldn't be left waiting too long in summer-time.

Cream-skimmers were used to lift off the cream. These worked well if they were shallow with a thin, almost sharp, edge. Skimmers from the last couple of centuries were often saucer-shaped with perforations to catch the cream while letting milk drip back into the pan, just like those used to remove surface "scum" from stock. Brass cream-skimmers on long or short handles are decorative antiques now, but some were much simpler. Anything the right shape would serve the purpose, like this wooden skimmer made to an older design. Other names for these were fleeter, scummer, skimming spoon, skimming ladle.

Churning should take place at least twice a week in summer.
Esther Copley, Cottage Comforts, England 1825

What is customer churn?

Customer churn, also known as customer attrition, is when someone chooses to stop using your products or services. In effect, it’s when a customer ceases to be a customer.

Customer churn is measured using customer churn rate. That’s the number of people who stopped being customers during a set period of time, such as a year, a month or a financial quarter. By expressing customer churn with a metric like this, you can turn it into like-for-like data that helps you measure progress over time. You can also express your churn rate in terms of dollar value if it makes sense to do so.

When defining churn, it’s important to be clear about when you consider somebody to have churned. Some sales cycles are longer than others. For example, in some industries, such as optical eyewear or home furnishings, it’s typical for customers to go for long periods without making a purchase because of the nature of the product, not because they’re under-engaged or at risk of churn. For each product and service you provide, fit your churn definition to your typical sales cycle, otherwise you may end up making reactivation efforts with customers prematurely.


Nearly any statistical model can be used for prediction purposes. Broadly speaking, there are two classes of predictive models: parametric and non-parametric. A third class, semi-parametric models, includes features of both. Parametric models make "specific assumptions with regard to one or more of the population parameters that characterize the underlying distribution(s)". [3] Non-parametric models "typically involve fewer assumptions of structure and distributional form [than parametric models] but usually contain strong assumptions about independencies". [4]

Uplift modelling Edit

Uplift modelling is a technique for modelling the change in probability caused by an action. Typically this is a marketing action such as an offer to buy a product, to use a product more or to re-sign a contract. For example, in a retention campaign you wish to predict the change in probability that a customer will remain a customer if they are contacted. A model of the change in probability allows the retention campaign to be targeted at those customers on whom the change in probability will be beneficial. This allows the retention programme to avoid triggering unnecessary churn or customer attrition without wasting money contacting people who would act anyway.

Archaeology Edit

Predictive modelling in archaeology gets its foundations from Gordon Willey's mid-fifties work in the Virú Valley of Peru. [5] Complete, intensive surveys were performed then covariability between cultural remains and natural features such as slope and vegetation were determined. Development of quantitative methods and a greater availability of applicable data led to growth of the discipline in the 1960s and by the late 1980s, substantial progress had been made by major land managers worldwide.

Generally, predictive modelling in archaeology is establishing statistically valid causal or covariable relationships between natural proxies such as soil types, elevation, slope, vegetation, proximity to water, geology, geomorphology, etc., and the presence of archaeological features. Through analysis of these quantifiable attributes from land that has undergone archaeological survey, sometimes the "archaeological sensitivity" of unsurveyed areas can be anticipated based on the natural proxies in those areas. Large land managers in the United States, such as the Bureau of Land Management (BLM), the Department of Defense (DOD), [6] [7] and numerous highway and parks agencies, have successfully employed this strategy. By using predictive modelling in their cultural resource management plans, they are capable of making more informed decisions when planning for activities that have the potential to require ground disturbance and subsequently affect archaeological sites.

Customer relationship management Edit

Predictive modelling is used extensively in analytical customer relationship management and data mining to produce customer-level models that describe the likelihood that a customer will take a particular action. The actions are usually sales, marketing and customer retention related.

For example, a large consumer organization such as a mobile telecommunications operator will have a set of predictive models for product cross-sell, product deep-sell (or upselling) and churn. It is also now more common for such an organization to have a model of savability using an uplift model. This predicts the likelihood that a customer can be saved at the end of a contract period (the change in churn probability) as opposed to the standard churn prediction model.

Auto insurance Edit

Predictive modelling is utilised in vehicle insurance to assign risk of incidents to policy holders from information obtained from policy holders. This is extensively employed in usage-based insurance solutions where predictive models utilise telemetry-based data to build a model of predictive risk for claim likelihood. [ citation needed ] Black-box auto insurance predictive models utilise GPS or accelerometer sensor input only. [ citation needed ] Some models include a wide range of predictive input beyond basic telemetry including advanced driving behaviour, independent crash records, road history, and user profiles to provide improved risk models. [ citation needed ]

Health care Edit

In 2009 Parkland Health & Hospital System began analyzing electronic medical records in order to use predictive modeling to help identify patients at high risk of readmission. Initially the hospital focused on patients with congestive heart failure, but the program has expanded to include patients with diabetes, acute myocardial infarction, and pneumonia. [8]

In 2018, Banerjee et al. [9] proposed a deep learning model—Probabilistic Prognostic Estimates of Survival in Metastatic Cancer Patients (PPES-Met)—for estimating short-term life expectancy (>3 months) of the patients by analyzing free-text clinical notes in the electronic medical record, while maintaining the temporal visit sequence. The model was trained on a large dataset (10,293 patients) and validated on a separated dataset (1818 patients). It achieved an area under the ROC (Receiver Operating Characteristic) curve of 0.89. To provide explain-ability, they developed an interactive graphical tool that may improve physician understanding of the basis for the model's predictions. The high accuracy and explain-ability of the PPES-Met model may enable the model to be used as a decision support tool to personalize metastatic cancer treatment and provide valuable assistance to physicians.

Algorithmic trading Edit

Predictive modeling in trading is a modeling process wherein the probability of an outcome is predicted using a set of predictor variables. Predictive models can be built for different assets like stocks, futures, currencies, commodities etc. [ citation needed ] Predictive modeling is still extensively used by trading firms to devise strategies and trade. It utilizes mathematically advanced software to evaluate indicators on price, volume, open interest and other historical data, to discover repeatable patterns. [10]

Lead tracking systems Edit

Predictive modelling gives lead generators a head start by forecasting data-driven outcomes for each potential campaign. This method saves time and exposes potential blind spots to help client make smarter decisions. [11]

Notable failures of predictive modeling Edit

Although not widely discussed by the mainstream predictive modeling community, predictive modeling is a methodology that has been widely used in the financial industry in the past and some of the major failures contributed to the financial crisis of 2007–2008. These failures exemplify the danger of relying exclusively on models that are essentially backward looking in nature. The following examples are by no mean a complete list:

1) Bond rating. S&P, Moody's and Fitch quantify the probability of default of bonds with discrete variables called rating. The rating can take on discrete values from AAA down to D. The rating is a predictor of the risk of default based on a variety of variables associated with the borrower and historical macroeconomic data. The rating agencies failed with their ratings on the US$600 billion mortgage backed Collateralized Debt Obligation (CDO) market. Almost the entire AAA sector (and the super-AAA sector, a new rating the rating agencies provided to represent super safe investment) of the CDO market defaulted or severely downgraded during 2008, many of which obtained their ratings less than just a year previously. [ citation needed ]

2) So far, no statistical models that attempt to predict equity market prices based on historical data are considered to consistently make correct predictions over the long term. One particularly memorable failure is that of Long Term Capital Management, a fund that hired highly qualified analysts, including a Nobel Memorial Prize in Economic Sciences winner, to develop a sophisticated statistical model that predicted the price spreads between different securities. The models produced impressive profits until a major debacle that caused the then Federal Reserve chairman Alan Greenspan to step in to broker a rescue plan by the Wall Street broker dealers in order to prevent a meltdown of the bond market. [ citation needed ]

1) History cannot always accurately predict the future. Using relations derived from historical data to predict the future implicitly assumes there are certain lasting conditions or constants in a complex system. This almost always leads to some imprecision when the system involves people. [ citation needed ]

2) The issue of unknown unknowns. In all data collection, the collector first defines the set of variables for which data is collected. However, no matter how extensive the collector considers his/her selection of the variables, there is always the possibility of new variables that have not been considered or even defined, yet are critical to the outcome. [ citation needed ]

3) Adversarial defeat of an algorithm. After an algorithm becomes an accepted standard of measurement, it can be taken advantage of by people who understand the algorithm and have the incentive to fool or manipulate the outcome. This is what happened to the CDO rating described above. The CDO dealers actively fulfilled the rating agencies' input to reach an AAA or super-AAA on the CDO they were issuing, by cleverly manipulating variables that were "unknown" to the rating agencies' "sophisticated" models. [ citation needed ]

Introduction to Churn Analysis

Image Source:

Churn Analysis which is also referred to as the Rate of Attrition can be defined as the process of analyzing data to understand why customers stopped using certain products or services. It can further be defined as the rate at which customers stop doing business with an entity or the rate at which employees leave their position in a firm. It doesn’t only give you the rate at which users stop using your services but also tells you why, when, and how to fix the problem.

It is not possible to change the strategy of your business because of a reduction in customer retention without carrying out a comprehensive study behind the reasons why they left, categories of those who left, and mapping out models to alter future occurrence.

Churn Analysis is also very important for a subscription-based business as it calculates the rate at which subscribers discontinue their subscription within a period and it is expressed in percentage. For such organizations to increase their client base, the Growth Rate which is the number of new acquisitions must exceed their Churn Rate. There are various widely used softwares that can help you perform Churn Analysis. Churn Analysis in Excel is considered to be the easiest to perform.

. And 5 Steps for Conducting Churn Analysis

Now that you understand the main reasons that cause customers to churn, here are the places you can look and the strategies you can utilize to conduct customer churn analysis across your business.

1. Evaluate Competitor Strengths and Weaknesses

Like we discussed earlier, if a competitor has a strictly better pricing structure than you while maintaining the same level of quality, then it’s a no-brainer for a customer to churn in favor of your competitor. Though things are rarely that simple, the principle is still the same. You should be constantly researching your competitors to identify areas where they are outperforming you and work to strengthen those aspects of your own business to reduce customer churn.

Not only should you be pinpointing competitor strengths, but also places where they are vulnerable and weaker in their offerings than you so you know where to take advantage of your own strengths and potentially bring their customers over to you.

2. Investigate Support Tickets for Repeat Issues

One of the primary aspects of your churn analysis strategy should be to identify any frequently-occurring customer complaints. Whether your customer service team uses support tickets, chatbots, or another way to track issues, this should be one of the first variables you investigate. These complaints come from customers who were upset and motivated enough to make the effort to reach out, indicating that there could be a significant shortcoming in your product that you weren’t aware of. You should be compiling and categorizing this feedback on a regular basis, and addressing any repeat issues ASAP.

3. Apply Sentiment Analysis Across Channels

To take the previous churn analysis strategy to the next level, you should be applying sentiment tracking and analysis across all of your channels, including customer support tickets, online reviews, social media mentions, and anywhere else customers communicate about your brand. By using AI-powered text analytics to compile customer feedback on a large scale, you can make more informed, data-backed decisions about what exactly is causing customers to churn. Learn more about Chattermill’s advanced CX tracking capabilities and how they can help your brand conduct customer churn analysis.

4. Discover Customer Drop-Off Points and Corresponding Changes

Even if users of your product aren’t leaving a specific review or contacting customer support about certain issues, you can use churn analysis to figure out if there were specific changes that caused them to leave. For example, a beachwear company might experience a significant churn in the off-season, as not many people are going to be shopping for a new swimsuit in the winter.

Or perhaps your company recently implemented a new AI chatbot, and at the same time saw a high level of customer churn. This could indicate that customers are having problems with the new technology. By pinpointing changes within the business and corresponding churn, you will be able to discover things about your brand that customers clearly don’t like, even if they’re not specifically saying so in reviews.

5. Utilize Customer Segmentation for More In-Depth Analysis

Not all customers are the same, so to make your churn analysis that much more insightful you should also be segmenting your customers by age, gender, income, subscription type (if applicable), history with your product, and any other relevant data points. This way, you can make sure that your customer outreach is as effective as possible and addressing the unique needs of each segment.

You don’t want to treat a brand new customer the same way as you would a 10-year, enterprise-level subscriber. Segmenting customers can also help you identify upsell opportunities if you are able to track a pattern for which types of customers are more likely to request additional features at specific times in the customer journey.

Take a look at the visual below for consolidated takeaways on what causes churn and ways you can conduct customer churn analysis:

Maintaining happy, loyal customers is the key to any successful business, and Chattermill’s advanced, AI-powered text analytics and sentiment analysis tools can help you conduct churn analysis to make sure your customers are staying satisfied. Contact Chattermill to set up a demo and see it in action.


If you have ideas or want to share your thoughts about our work, our data science team would love to hear from you. Please send feedback to [email protected]

Meet the Moguls: 5 Takeaways from Emily Ley, Barbara Corcoran, and Andrew Bialecki

Why Brands Can’t Pin Their Hopes on Prime Day

A CMO’s Advice on How to Deal with Changes to Consumer Data Privacy

Get email marketing insights delivered straight to your inbox.

Christina Dedrick

Christina is a data scientist at Klaviyo. She’s helped build to several of Klaviyo’s data science products: Benchmarks, Smart Send Time, Predicted Gender, and CLV. When not creating tools to make it easy for businesses to leverage their data, Christina can usually be found hiking or rock climbing.

Making butter is an ancient practice that is even noted in scriptures.

The concept of separating cream from fresh whole milk and shaking or churning it into butter is well documented in ancient history.

The discovery of churning butter later lead to the discovery of making cheese.

Sticks of butter as we know it today were not sold in some rural area stores until the mid 1900s.

Butter was either churned at home or sold directly by a dairy. Some of these suppliers were also called creameries.

Up until the mid 1960s butter was still being hand churned on most rural farms.

Most of these farms were very self sufficient and either owned a milk cow or traded other homegrown commodities for fresh milk and cream.

Even today butter is still being hand churned or churned in small electric butter churns in many rural homes.

Once the butter was made it was taken out of the churn, placed into a bowl and rinsed with cool water until the water ran off out clean.

The butter was then worked on a board until all the water was removed.

The finished product was then scooped into a dish or wooden butter mold to be shaped and formed.

Many of the restaurants and fancy hotels churned their own butter and placed it into wooden individual patty signature molds to display their logo or symbol.

Buttermilk is a result of the left over remaining milk left in the churn.

Buttermilk is a favorite ingredient used by of cooks to enhance many of their baking recipes.

Some might even confide that it&aposs their secret recipe.

The History of making butter

With little documentation of the actual discovery of making butter the assumption prevails that butter was first discovered by nomad herders who carried some milk in leather saddle bags.

At the end of a journey they discovered the milk had been shaken so bad that it had turned to a solid matter with a different tasting milk. This milk would become known as butter milk.

Butter was once considered a lavish dish suited for royalty. The delicious discovery did not last very long as it quickly had to be used up before it would turn rancid.

The Celtic people are reportedly the first to use salt to preserve butter for a longer table life.

The history of butter churns

The first known butter churns still used leather pouches that were hung and shaken by different manual means.

Apparently at some point someone discovered the same result occurred in shaking milk in a wooden barrel.

Further discovery revealed that when the barrel was outfitted with wooden vanes on the inside the milk could be made into butter by rolling it.

It is unclear of when the plunger that would become known as the �sher” was first used.

The wooden butter churn with dasher known as a �sh churn” became the household source of converting cream into butter.

The wooden dash churn is still being used in many European countries.

The rolling wooden barrel was also placed on a stand with a hand crank and then later was developed as a stationary wooden barrel with a hand crank attached to an inside wooden impeller.

At one point even a wooden barrel was attached on each side of a stationary base with a rotating hand crank that turned the barrel from end to end.

It is evident by the early discover of the wooden vanes inside the wooden barrel that the cream needed to splash against something to begin the transition into butter.

Without a splash bar the cream would simply roll around in the barrel.

Many people believed witchcraft was at play if their cream didn&apost churn into butter.

Several patents took the wooden barrel churn and replaced it with a square or octagon box so the cream couldn&apost just spin inside.

Documentation does not really begin to show the steady and rapid developments in the butter churns until people began to patented them in the 1800s.

Through the late 1800s and up until the mid 1900s over 2500 patents were assigned to different butter churn designs.

Evidently gaining a patent during in this era was very easy even with the very close similarities of other issued patents.

In looking over the glass butter churns discussed on down in this hub, side by side there is very little difference in the design and functioning of the hand crank models, yet so many different patents were issued for them.

Several patents took the wooden barrel churn and replaced it with a square or octagon box so the cream couldn&apost just spin inside.

Patents were also issued for a rocking churn that used variations of a wooden box with baffles on the inside to churn the cream as it rocked across the baffles.

Other designs took a stationary wooden box and equipped with it rocking baffles attached to a handle to rock the cream back and forth.

As more people worked on the development of making churns to produce butter faster, others worked on the transition of a leak proof container.

As people developed hand glazed pottery bowls, crocks and dishes the transition over to a glazed pottery dash churn also came out of the 1800s and became the choice up into the 1900s.

Little was known about bacteria and sanitary methods in the 1800s but the choice of a ceramic churn base would have become significant as that knowledge grew.

Tin was also used in butter churns. In review of the older barrel and rocking churns tin is noted as the outside skin in place of wood on several different models.

A tin box base was also used on the hand crank box churns that came out of the many patents filed.

One such patent for a tin butter churn was issued to a gentleman by the name of Nathan Dazey who latter founded the Dazey Churn Company.

The progression of the glass butter churns in America

As evidenced by the amount of butter churns available in antique stores today, Dazey was probably one of the most successful butter churn manufacturers or maybe best marketers?

Nathan Dazey of Dallas held the patent to the tin box butter churn, which became widely used up until they began selling glass butter churns beginning in 1922.

The name Dazey would later dominate the butter churn market in the US with its version of the glass jar churn using the hand crank technology they had developed so well with the tin churns.

The hand crank tin churn utilized a gear system that turned a rod fitted with wooden or metal paddles.

The very first glass jar churns were actually developed in England by the Blow Churn Company beginning in 1900 and were made up until 1929.

From that point the transition of the glass butter churn into America gets a bit fuzzy as there seems to be several different accounts of how and when the glass churn began being manufactured in the US.

Interesting the glass churn designed jumped the pond and was granted a US patent in 1921.

It’s also unclear how the transition occurred from England to the US, it appears that it was a collaborative effort with the US patent holder.

Soon after the patent Taylor Brothers Churn Company of St. Louis Mo. began manufacturing a glass butter churn very similar to the Blow Churn under the name of �ndy”

Prior to the glass churn Nathan Dazey who was making can openers in Dallas Texas purchased the EZ Butter Churn Company.

This purchase provided the path for Dasey to enter the butter churn business starting with the tin butter churn that he patented in 1917.

With his success with the tin butter churn Nathan Dazey became intrigued with the idea of making a small home style churn for the small family rural farm.

Most of these farms had a daily supply of fresh whole milk but it was a limited supply from one cow.

The larger sized tin churns were difficult for the small family farm because electricity was at the time not available and refrigeration was very limited.

Milk and cream had to be processed quickly and could not be held over long to make larger batches of butter.

With his hand crank technology Dazey figured he could easily adapt it to a smaller glass jar similar to the Taylor Brothers Churn.

The Dazey Manufacturing Company grew to the point that his son John joined forces with him and they moved the company to St. Louis Mo.

It appears some collaboration may have also occurred with the St. Louis based Taylor Brothers Churn Company, but Dazey was granted a patent for his very similar glass churn in 1922.

Later reports indicate the Dazey purchased the Taylor company in 1945.

It’s also interesting that Dazey purchased the Standard Churn Company in 1950 and began using the Standard Churn hand crank design, the metal blades and even the tulip glass jars.

Apparently Dazey felt that the Standard Churn was a better design even after all the years of success with the original Dazey churns.

So is it just that Dazey was better at marketing, or was Standard beginning to out perform Dazey in quality?

At that time news traveled slow and people didn&apost have the sources of product reviews to know what was best. They could only rely on the advertisements of the manufactures.

Dazey already had solid name recognition through it&aposs established tin churns, which probably gave them the leg up over Taylor and Standard.

If you have a Dazey churn with the metal blades and tulip glass jar it probably was made post 1950.

Several other companies made glass butter churns, Elgin and Fulton were two such manufactures.

These companies also plain labeled churns for such mega mail order companies as Sears Roebuck and Montgomery Ward.

Sears and Wards were the early day discount stores which gave Dazey a hard run in the market place.

This period during the 1930s was especially difficult for companies trying to sell new products as the country was still trying to recover from the depression.

Dazey came out with an economy churn called the Price Churn in an attempt to keep the market share.

Other companies made their own churns or contracted with a churn company to plain label churns for them.

We found a couple glass jar churns as the one pictured that only had a simple wooden plunger dasher.

We couldn&apost ID a manufacturer or find any similar product on the Internet for this plunger churn.

The plunger appears the same hand held wooden mixer that we have seen in other antique stores.

They apparently were used by rolling the wood dowel handle back and forth between the hands while plunging up and down into the glass jar.

The plunger churn appears to be factory made but could have been very well homemade or made by a small regional manufacturer in Texas.

Texas is the only place that we have ever seen this wooden plunger churn.

The other possibility is what we have seen in the other antique stores may have actually been the wooden dasher salvaged out of one of these glass jar plunger churns.

Sunbeam also had a glass jar churn on the market as an attachment to their popular Sunbeam Mix Master electric mixer.

We were not aware of the Sunbeam mixer attachment until we stumbled across it in an Antique store and just had to have it.

Regardless of the many different churns made during this era, Dazey seemed to dominate the butter churn market with both their tin and glass jar churns.

This is one of the reasons that Dazey churns are still so available in antique stores.

The Dazey glass churn became popular with families who wanted to make quick small batches of butter.

It’s size allowed it to be easily stored out of the way on a pantry shelf.

The Dazey tin churn in the larger sizes remained popular with the commercial size dairy operations.

As Dazey perfected and modernized the churn the open hand crank was enclosed with a more modern gear box that some call the red foot ball design.

Dazey continued with innovations with a strainer screen in the top and then the electric motor versions as the rural areas became electrified.

The hand crank versions were still used up into the 1960s as many rural homes didn’t obtain electricity until the mid 1960s.

Dazey hand crank churns are still being used and even reproductions of a hand crank churn are being used on rural farms and within sustainable living cultures.

These reproductions are actual working butter churns. Since Dazey churns have become such popular collector items fetching around $175 each many other antique reproduction Dazey churns have entered the market place.

Dazey never made a churn smaller than a 1 qt so the small so called salesman sample is for sure a fake.

Other butter churns used today for the home and small commercial dairy operations are made of Stainless Steel.

If you’re interested in locating a working reproduction churn to actually make butter, online self sufficient back to basic stores like Cottage Craft Works .com carry new glass, pottery, and stainless butter churns.

2. Development

Zero hour

Ok, so let’s sum up what we have. We have telecoms overwhelmed with the amount of data and irritated customers. The telecoms struggle with decreasing customer retention and growing churn, while customers don’t get the offers they seek.

Being a problem of most telecoms worldwide, growing churn was also a problem of one of the major Polish telecoms when they approached Neoteric . With the help of SaaS Manager, we intended to help them increase customer retention. Our major goals were to:

  • implement AI and predictive analysis to better understand clients’ preferences , which was measured with the churn rate (we wanted to reduce it by 2 percentage points in one year from the product launch),
  • transfer our knowledge so that the Client’s analysts would be able to build and use their own predictive models.

Solution – how we relieved some of these pains in one of the major Polish telecoms

Step 1: Prototype

After identifying the Client’s main pain points, we started the project with prototyping. The prototype we made was a sample model predicting churn . Its aim was to show that with specific data, it will be able to predict which customers want to leave, changing their service provider or resigning from the offered service for good.

The model was trained using some sample data and it was available as a service with an interface for integration. As the customers’ data processed by the model was shared through the API, it was easy to process it further – regardless of the tool that our Client’s team might want to use or will use in the future.

Step 2: More predictive models

The next step was to introduce another model and to merge it with the one predicting churn. The new model was meant to assess customers’ buying preferences. Once a customer was identified as one who was likely to leave, the model would suggest the consultant what products they should present to increase the chance of retention. The idea was to offer customers only the products that would fit their needs.

Once the two models were ready and successfully merged, it was the time to use some real data collected by the company. We initially worked with the Customer Retention Department, planning to start the pilot after 9 months from the beginning of building predictive models. Working Agile, we quickly discovered what features are really needed to start testing the models in action. After reviewing the initial plan and verifying some of the basic assumptions, we were ready to launch the pilot after less than 3 months!

That was the time when we were able to expand our activity and work with different departments. After achieving the set goals in the most demanding area of the market, we expanded our predictive models to work on the whole database and added some tools to support the Sales and Marketing Departments.

One of the most important challenges of this stage was to make sure that data records of millions of the Telecom clients are secure and no unauthorized person has access to this data – and we achieved both goals by applying proper data anonymization procedures.

As we dealt with sensitive personal data, it was very important to secure all the records from the very beginning. For us, as a third-party contractor, every end customer had to be anonymous. To ensure that, we implemented a solution that generated one-sided hashes based on customers’ data. From our perspective customer was a long identifier like “XXXX-YYYY-ZZZZ-1234” not Mr Jan Kowalski living in Warsaw, without possibility of reversing such hashing to get to know the data of the actual person. When someone from the Client’s team needed to fetch that person scores, he was able to send the request that was generating hash on the fly to get to look up a profile of desired customer – and to receive the right record back. To prevent potential data leaks, we made sure that each time someone requests the data about the specific user, his or her name was stored in the access history. If any data was exposed, it would be easy to track who was responsible for the leak.

Data anonymization was also crucial to comply with the GDPR policy. As every piece of available data was anonymous, we were able to use it to train the models. The outcome, however, was applied only to those customers, who agreed for the profiling – only those ones would receive personalized offers based on their behavior and the predictions about their future choices.

At this stage, we delivered a trained predictive model predicting the risk of churn and assessing clients’ buying preferences as well as all the mechanism that kept churn scores updated. As the model was trained on real, though anonymized, data provided by the Client, it was ready to test.

Step 3: More data

The next challenge was to make the predictions more precise and to make sure that the model is able to handle big data. At the previous stages, the predictive models were training on a part of the available data. Later, we could introduce data from various systems.

Connected to an up-to-date database and API with the scoring, the models were ready to release their full potential. The database consisted of around 150 million of records related to customers and another 75 million of additional, external metrics. Our models are able to process and update the whole database in less than 24 hours! And the scoring of the whole customers’ base lasts less than one hour. Can you imagine how long it would take a human being?

After processing the data, the predictive models could present the Sales Team the new scoring information which showed the probability of resignation along with the probability of interest in specific products . This meant that the consultant would get a report showing that Mr. John Doe is likely to change his service provider and that if he stayed, he would be interested in sport channels at 40% and in comedy channels at 60%. This way, the consultant would be able to call Mr. Doe and offer him something that may actually stop him from leaving.

As the next step, we started to work on the scoring system that would show the complex offers and organize them from the most probable to satisfy the client to the least probable. With the choice of hundreds of different (and regularly changing) offers, it would be impossible for any consultant to know all the offers, not to mention choosing the best one for each client.

Our models did not only analyze the behavior of the chosen subscriber, but also the behaviors of thousands of other subscribers – which led to creating behavior patterns . Thanks to that, the system was able to suggest offers that were personalized and adjusted to the client’s current needs.

Step 4: Pilot

That was the time for trial by fire. In order to check the performance of the designed solution and to implement further improvements, we decided to test our churn models on the full customer base. Working with the Sales Team, we aimed to decrease churn by 2 percentage points per year , but working only with the first, most difficult, most demanding segment of customers, we’ve managed to exceed this value!

The predictive models we created with SaaS Manager were able to analyze the behaviors of the clients, predict when they may want to leave and suggest the offer that can stop them. We analyzed over 200 million of records and ran the first experiments. From October 2017 to April 2018, we contributed to decreasing the number of resignations by more than twice of what was expected from us*.

The models are also able to predict when the customer may want to change their service provider and what the consultants can do to stop them. The consultant will not bother the customer before it’s needed, which creates a win-win situation: the customer is not annoyed with intrusive calls and the consultant does not waste their time on reaching customers who don’t need it .

* The exact number has to be disclosed due to NDA.

Step 5: Know-how

The final step was to transfer our know-how. As we’ve already delivered tools to do it, it was important to teach the Client’s team about creating such models so they could later train the created models or build similar ones by themselves.

We provided the Client with:

  • a guide to creating predictive models,
  • tools to build and train predictive models,
  • Integrated Client Profile – an up-to-date database of the Client’s customers with their scoring
  • trained models that predict buying preferences of our Client’s customers and the risk of their churn,
  • Scoring Pipeline – set of tools to continuously update profile data and scores, including new models built by our customer analysts

With this knowledge and the toolset, the Client’s team was able to use the data provided by the predictive models we’ve build and to build and train new models . So far, the Telecom team was able to develop sales predictive models for the new offers and merge them with the ones created by our team. Another model they have built helps them predict possible breakdowns of the infrastructure and to plan the expenses related to its maintenance.

During the pilot, we were able to teach them good practices related to project management: Agile, DevOps, using clear KPIs, closely aligning with the Client’s team, and handling regular meetings – helping the Client’s team work more effectively.

Building a Model

Creating a great dataset is the hard part. With no-code tools like Apteo, building a churn model is easy.

First, connect your dataset. Below, I simply drag-and-drop a CSV file of my churn data into the platform. Then, I head to the “Predictive Insights” tab and select “Churn” as my KPI. I leave the default settings as they are, and an automated machine learning model gets created in the background.

Now, I can see how different attributes impact churn, and I can predict whether a customer will churn by putting in data like their monthly charge and tenure.

Video, Sitemap-Video, Sitemap-Videos