Price prediction in Healthcare Tendering

Predicting the final price of an auction has become an increasingly important area of business improvement for healthcare providers. Manufacturers can obtain a large profit by using previously predicted prices. 

At Konplik Health, we have over 6 years of experience delivering successful tendering projects for leading healthcare companies in more than 40 countries. From this, we can certainly voice that, at tendering opportunities in which price is a key criterion, both manufacturers and service providers in healthcare can increase their margins and revenues through predictive analytics. 

Predictive analytics uses statistics, advanced mathematics, and artificial intelligence to study large amounts of current data and past events so, patterns can be identified that allow forecasts to be generated for future events. It uses complex analytical models to analyze patterns in raw data and in turn, predict future results

Originally published at https://konplik.health/price-prediction-in-healthcare-tendering/


Automation and AI working together for tendering

Healthcare tendering is a challenging task for manufacturers of medicines, medical supplies, and equipment. 

The diagram below describes the steps we follow to deliver price-predictions at tendering. 

Benefits of process automation: data gathering, extraction, transformation and loading


On one hand, information gathering and managing processes can take weeks or months to complete. Thanks to advances in automation, however, a data-centric approach can often replace manual or semi-manual processes. 

What machines can automatically do:

  1. Data gathering

-Gather structured and unstructured information via scraping. 

  1. ETL (extraction, transformation and loading)

-Convert raw sources into a structured or semi-structured format.

-Combine public and private data

-Transform data with custom filtering, fuzzy product matching on large sets of data.

Modeling in price prediction

On the other hand, the lowest price is the key criterion for selecting the winning supplier at many healthcare tenders. Tender-based purchases are implemented to minimize the price for the duration of a contract. This practice is expected to reduce the cost as a result of price competition. 

Margins, as a result, remain under pressure due to the competing agents that bring about a never-ending price erosion. 

At tender opportunities that favor the lowest price, providers are required to submit a quotation. That is a fixed, non-negotiable price offer that cannot be changed once accepted. 

By collecting and analyzing past performance data from tenders, as well as current market prices and risk assessments, providers can take on a data-driven approach to bid more efficiently on market opportunities.

Creating predictive models requires more steps. As the diagram above shows, the train-test split is a technique for evaluating the performance of a machine learning algorithm. The procedure involves dividing a set of data into two subsets. The first subset is used to fit the model and is called the training data set. Contrarily, the second subset is not used to train the model.

One model doesn’t fit all

Not all prediction models relevant to prices in health tenders can be transferred from one market to another.

Despite the success of certain models in some markets, it is still incredibly important to be skeptical and question assumptions about the data and the analysis methods used.

Each tendering space is determined by different conditions that alter human decision making. One predictive analytics solution will not fit all. Empirical data proves this time and time again, as it frequently refutes carefully woven theories and models that work in other markets. 

In an upcoming post, we insist on a critical approach towards the blind application of game theory in predictive models of health tenders. 

Each market has its own rules for contracting, varies in the number and aggressiveness of competitors, and the availability of information from the players. As a result, our predictive models radically change from one market to another in order to obtain the best results.

We regularly test several models to find the best fit for each market behavior. 

Available information shapes competition

In our experience, the information available to competitors is the most dominant factor in shaping their decision-making.

In some markets, competitors have frequent and up-to-date information on bidding results. In Scandinavian markets, for example, healthcare tenders take place once or twice a month and competitors have precise information on how each player bid in previous opportunities. These conditions considerably limit the range of variability in bids placed by players.

In other markets, little or obscure information is the general rule.

In England or Wales, for example, the NHS intentionally withholds relevant information from competitors, so that their bids accelerate price erosion.

When a competitor loses in a bid, they are only informed of their approximate distance from the winning bid, and not the exact price that the bid won at. Likewise, when a bidder wins, they are also only given their approximate distance from the second-place bidder, precisely to get an even more moderate price in the next opportunity.

Evidence-based algorithms

The good news is that, while prediction models cannot always be applied to other contexts, methods often can.

We have successfully deployed tailor-made algorithms (including chunks of game-theory) along with machine learning algorithms (including gradient boosting, random forest, and Bayesian networks).

The key to obtaining a profit using previously predicted prices is to constantly check and review models with respect to the results obtained. 

On a regular basis, our algorithms automatically pitch a number of models against each other in simulations that decide which model was the fittest in the past, in order to subsequently decide which one to apply for the next bidding of a given product. 

For example, in some healthcare markets a buyer will purchase close to 2000 products at each tendering opportunity. The manufacturer we work for might bid for up to 50 products. For each of the products that are contracted through tender, we carry out a simulation to determine which algorithm would have obtained the best results in the past. 



Examples of algorithms

In the simulated competitions between algorithms, we have pitched algorithms such as: 

  • Tailor-made game theory-based algorithms. Stay tuned for a new post on this subject soon. 
  • Gradient boosting algorithms. Also, tailor-made. 
  • Tailor-made Time Series Models with some customization
  • Ridge Regression
  • Classification/Decision Trees
  • Random Forest
  • Support Vector Machine (SVM) for regression analysis
  • Bayesian networks
  • Neural Networks


The model that has obtained the best results for a given product, (this could be gradient boosting, for example), is the one we apply the next time to recommend a price for our client. A new simulated competition is carried out for each new product. This time a different model (Random Forest, for example) might prove to be the fittest in the simulation, and, as result, be the one that we deploy for the second product. 


Metrics used at price prediction

We regularly assess the performance of our algorithms using a variety of metrics that include

  • Margin: revenue minus the cost of goods
  • Revenue growth
  • Money left on the table: The difference between the submitted bid and the lowest-priced competing bids
  • Win ratio


Despite the success of certain models in some markets, we can say that each tendering space is different, and accordingly, one model will not fit all. Empirical data analysis is required to be able to increase the margin contribution and revenue growth through predictive analytics. 


Sales through tenders in healthcare are becoming a vital element in the strategy of manufacturers. Healthcare tenders affect purchases of medical supplies and equipment, purchases of medicines (especially for non-patent pharmaceuticals), and services.

Healthcare tendering can be defined as: “The bulk purchase of goods and services through a competitive bidding process.”

Originally posted at https://konplik.health/tendering-strategy-in-healthcare/

Often, tender-based purchases are implemented to minimize the price for the duration of a contract. This practice is expected to reduce the cost as a result of price competition. It regularly achieves economies of scale and can also reduce some inefficiencies caused by segmented distribution systems.

By some estimates (3), around 25% of the turnover in Pharma is currently coming from tenders. Pfizer, for example, has a network of more than 80 countries planning, managing, and tracking tenders (3).

Bidding for healthcare products is a common practice in both developed and emerging countries (1-2). Healthcare companies operating in global markets know that tenders are an inescapable reality wherever price sensitivity is an essential factor in the purchasing criteria.

The drive to use bidding is growing in Western Europe (1), even with its widespread tradition of publicly funded health systems.

Key markets, such as China, Brazil, and Russia, are making progress in expanding health coverage, which will likely generate greater demand for public contracting and cost-saving strategies such as tendering. According to the same research by McKinsey, in the Asia-Pacific region, almost 51% of the population will have some form of universal health coverage in the next ten years.


Successful Tender management is challenging

 The management of tenders for healthcare goods and services, however, involves significant challenges.

The types of tenders can be quite varied. Seemingly, the main types of bidding processes are:

  • Open tender, which is preferred in a public tender.
  • Tender, which is restricted to those who have previously proven to be qualified suppliers.
  • Negotiated tender

Assuming the suitability of the bidders, open and restricted tender have the following procedure:

  1. Public notice of the specification by the procuring entity
  2. Presentation of offers by bidders
  3. Evaluation of offers
  4. The award

The negotiated tender follows the same procedure for the most part; but, instead of carrying out a mere evaluation of the bids, the procuring entity enters into negotiations with bidders on their initial bids.

On the other hand, contracting authorities often establish strict criteria in terms of competence, professionalism, and qualifications that must be met for a provider to be successful.

Even though the lowest price is the key criterion for selecting the winning supplier, some countries may take other elements into account, such as the ability to supply in quantity, supplier quality, or local manufacturing. From market to market, levels of competition and pricing tendencies may vary

High-quality data sources and analytics

To obtain profitability a company must have access to timely and reliable sources of high-quality data, that allow them to make better decisions.

First of all, providers need to gather and structure vast amounts of content published and managed on public websites.

Due to the volume of information about new tender opportunities and awards worldwide, it is not efficient to employ people to conduct the repetitive task of browsing through web pages.

This is where automated data crawling services become invaluable. It plays an important role in giving organizations a significant advantage over their competitors.

Experts who are responsible for participating in tenders require timely business intelligence with data that allows them to make better decisions.

Once the data has been extracted from the information sources, it is necessary to structure it using natural language processing tools.

Next, it is necessary to combine public sources with private company data.

For example, an automated tool is able to recognize the name of a molecule or a medical device, which often comes in many different shapes and forms (even with typographical errors), and link this to a product that the supplier manufactures and can sell.

Predictive analytics looks at patterns in data to make predictions about future events based on current and historical data.

Automated algorithms can help us to carry out value-adding tasks with a significant edge over human experts. Such tasks include:

Price forecasting: Outputs a prediction of the winning price at tendering opportunities.

Demand forecasting: Returns the total demand for a given product in each market. How many tendering opportunities will happen in the next year (or quarter) for a given product? How many will we win? How much will we sell? At which price?

Revenue and margin forecasting: Used to predict and optimize revenue and margins, as well as expand the discovery of new business opportunities.

To summarize, although sales through tenders are becoming an increasingly vital element in the strategy of healthcare manufacturers, successful tender management remains challenging. To obtain profitability a company must have access to timely and reliable sources of high-quality data that allow them to make better decisions



1. Leopold C, Habl C, Vogler S. Tendering of pharmaceuticals in EU Member States and EEA countries: results from the country survey. 2008. Available from: https://ppri.goeg.at/sites/ppri.goeg.at/files/inline-files/Final_Report_Tendering_June_08_7.pdf

2. Drug tendering: drug supply and shortage implications for the uptake of biosimilars



Voice of Employee Dashboard Howto

In this step-by-step tutorial posted initially at MeaningCloud, I am showing to you how to create a nifty dashboard with Excel for the Voice of the Employee.
Voice of the Employee gathers the needs, wishes, hopes, and preferences of all employees within an organization. The VoE takes into account both explicit needs, such as salaries, career, health, and retirement, as well as tacit needs such as job satisfaction and the respect of co-workers and supervisors. This post follows the line of Voice of the Customer in Excel: creating a dashboard. We are creating another dashboard, this time for the Voice of the Employee.
Text-based data sources are a key factor for any organization that wants to understand the “whys”.
Things can often go wrong in organizations. It is important to catch those potentially damaging problems before they turn into a major crisis.
VoE programs typically use a variety of methods to collect and analyze employee feedback, including surveys such as eNPS or 360º assessments, exit interviews, employee forums, social networks or focus groups.

Segmentation: Divide and Conquer

Brian: You’re all different!
The Crowd: Yes, we ARE all different!

Monty Python’s Life of Brian
Can you segment employees as we do with customers so that activities like talent development improve and correspond to each employee’s profile?
In this post, I show you in a very practical example how we’ve brought a classic marketing technique to the people analytics environment.
At the end of the post, don’t miss the takeaways of the What is in it for me section.
RFM analysis, short for Recency, Frequency and Monetary value, is one of the customer segmentation methods that is easiest to deploy and, at the same time, returns the best results. We are showing how to apply it to the “internal customers,” a.k.a the employees of an organization.
In this case, we’ve applied the analysis to real estate agents, in order to implement customized training and economic compensation plans according to each agent’s performance.
Read on at AnalyticsinHr:


Six methods to measure performance

Using data to evaluate performance allows incentives to be more closely linked with encouraged behaviors.
In this post:
1.   Three subjective and three objective methods to measure performance.
2.   An actual performance evaluation case based on the 180° feedback survey methodology.
3.   A proposal to evaluate performance more frequently, using objective measures.

If you can't measure it...

Surely you've heard someone say before:
If you can’t measure it, you can’t manage it
Although I fundamentally agree, this quote is falsely attributed to many people, which commonly happens with quotes like this.
As a critical thinking exercise, I'm going to try to get to the bottom of this ultra-famous quote's origin.
The quote is usually attributed to Peter Drucker. If you search in Google Images, you'll find a bunch of photos of Drucker with the quote stuck to him, as if it were coming out of his mouth:

(July 15, 2017) http://bit.ly/2tS0YWO
There's a small problem. Peter Drucker never said it.
The truth is that in my tiny 30-minute search, I wasn't able to find out who said it first. Ed Sekyota appears among the candidates, but the quote I found was from 2003.
This sentence is also attributed to Edwards Deming. The problem here is that Deming said exactly the opposite. His quote is:
It is wrong to suppose that if you can’t measure it, you can’t manage it – a costly myth.
Dr. Deming wrote a lot about the value of using data to help improve managing an organization.
His phrase is really:
In God we trust, all others must bring data.
You can fact check me on the Deming Institute web page: (http://quotes.deming.org/category/data.
Deming also cautioned that you can't measure every important thing, but regardless, you have to manage those "immeasurables."
My search for the origin of this quote ends here. In the post-truth world we live in, its authorship is insignificant compared with the excitement that this unwavering truth fills us with.

Performance measurement

Good performance can be objective (sales, billable hours) or clearly subjective, like 360° surveys.

Subjective performance measurement

Let's look at the most important methods.

1. 360° surveys

360° surveys are definitely the most popular tool to measure employee performance.
Studies suggest that more than a third of American companies use some type of multiple evaluation. Others state that 90% of Fortune 500 companies apply some type of 360° survey.
However, there's no lack of criticism about these kinds of evaluations.
To evaluate an employee's score, their colleagues, subordinates, customers, and manager are asked to give ratings on specific topics.
If you're not familiar with the lingo, the degrees of the 360º are distributed like this:
90º = Evaluation of direct managers
180º = Manager + Colleagues (peers)
270º = Managers + Colleagues + Subordinates
360º = Managers + Colleagues + Subordinates + Customers

2.  Net Promoter Score

More and more, companies understand that increasing profit is very good, but in extremely competitive environments, they need to have satisfied customers. They also know that their employees have a fundamental role in improving the customer experience.
In terms of experience, the line between consumers and employees is actually blurring. The new goal in the best organizations is to create brands that deliver a unique experience for current and potential employees and consumers.
Hundreds of the biggest companies in the world like General Electric, Procter & Gamble, Telefónica, Apple, Allianz, JPMorgan Chase & Co., ING, and Microsoft adopted the NPS as the principal metric in measuring customer satisfaction. They have proven that growth and profits are directly related to converting more customers into promoters of their brand.
According to Wikipedia, in 2016, two-thirds of the Fortune 1000 companies have adopted this metric.
Along that same line, the mission of having motivated and happy employees also has its metric: the eNPS (employee NPS). I explained it in much more detail in Chapter 9.

3.   Forced ranking

Forced ranking, vitality curve, or popularly "rank and yank," is a very controversial management tool that uses very intense annual evaluations to identify employees with the best and worst performance through person-to-person comparisons.
Managers classify workers into three categories: the highest 20% are the "A" players, the people who will lead the company's future. They're given raises, stock options, and training. The 70% in the middle are the "Bs," the stable ones, who get smaller raises and are encouraged to improve. The 10% on the bottom are the "Cs," the ones who contribute the least. They don't get raises or bonuses, are offered more training, encouraged to look for another position, or they are fired.
General Electric, who fires 10% of its "C Players" every year, is probably the most well-known representative of this practice. Ex-CEO Jack Welch suggests that forced ranking helped increase GE's revenue to $130 billion in 2000, from $70 billion in 1995.


Up in the air

When talking about firings, the first thing that comes to mind is Up in the Air. Ryan Bingham (George Clooney) works for an HR consulting company that specializes in firing people. He makes a living traveling the United States firing people on behalf of the bosses of each company.
In the movie, Ryan's assistant, Natalie Keener, a young, ambitious 23-year-old screams...

Please, for the love of God, can I fire the next one?

Objective performance measurement

4. Sales

Sales are definitely the simplest way to measure performance. For obvious reasons, it can only be applied in departments where there is commercial activity.
Short sales cycles are the ideal territory for this metric.
However, when sales have a longer cycle like they often do in People Analytics sales projects, where you may need more than six months to sell a project; it's often necessary to use other additional metrics to evaluate the process. In general, you divide the sales funnel into phases and measure the interventions in each phase of the process: generating new contacts, phone calls, follow-ups, conversations, and sales.

5. Number of units processed

Think of a supermarket. How many products does a cashier scan in one hour of work?

Bloomberg, for example, counts the number of times that its 2,400 journalists press a key when writing their articles. Many companies measure the number of lines a programmer codes.
But there are a lot more: the number of web pages visited, instant messages sent, e-mails sent, open documents, log-ins, etc.

How many errors or how much they contribute

It's evident that measuring the processed units without looking at the quality indicators (how many errors made, how customers rate their experience) can end up being clearly insufficient or even deceptive.
- Think of a financial entity's team that reviews fraud cases. If they only count the number of evaluated cases, you have a deceptive performance measure.
A better way of measuring performance in fraud analysis is quantifying the economic impact of the intervention, that is, how much fraudulent money was detected? Looking at historical data, you calculate the average detected fraud before the specialized team's intervention. You measure again after the team puts anti-fraud activities into place. That way you have an actual contribution.

6. Lifetime Value

There are several studies on the average contribution per employee in tech companies.

Annual assessment vs. frequent assessment: transparency

Regarding annual or biannual evaluation, frequent performance scoring gives employees clear feedback about whether their performance is increasing or decreasing.
With this short feedback cycle, managers can reward hard work or rapidly take corrective measures if need be.
General Electric stopped using its legendary vitality curve ten years ago, the one that made them fire 10% of the worst of their staff of 300,000 employees.
GE is now introducing a new system that anticipates more frequent updates of its performance evaluations through an app.
With this decision, GE joins other important companies like Microsoft, Accenture, and Adobe, that have adopted a new focus that's based on more frequent review and feedback of employee performance.