66% increase in Revenue per sales call

66% increase in Revenue per sales call

A recognized European automotive trading company faced a typical challenge – how to get more cars sold to B2B customers with the limited size and capacity of their sales force team?

Goal setting
Each active customer was supposed to be contacted at least once a month. The best customers once a week. The better the customer, the more frequent contact was required.
With thousands of customers it was unrealistic for the ordinary sales reps to assess when to call whom. They needed a system to tell them, because in practice they made tens of calls daily, but more or less by intuition.

Behavioral prioritization
We believed that we could create a guidance system that would assess individual customers, sales reps and calls in order to bring-up a list of customers to be contacted in the near future preferably.
A customer scoring model leveraging recency, frequency and monetary value of their past purchases has been implemented and engaged. This way we could sort all customers by their importance, respective of their propensity to make another purchase in the near future.
As long as the Call Duration Records were available, we could determine when each customer had been last contacted and suggest the next call date based on the required contact frequency – of course within business hours.

Product recommendation
To increase the relevancy of offerings during the call we also trained a predictive model on the past purchases, which returns cars that the particular customer is most likely to order. Sales reps then stopped offering just “something”, they rather started to let the customer choose from a few, very relevant, options.
When the management saw it, they let us implement that algorithm also in the company’s online e-commerce portal.

Stock-aging and optimization
Many of the cars traded are being sold before getting produced. The rest of them get sold after being produced – which means that they get physically delivered and stocked.
For the company it is desired to shorten the sales cycle as much as possible. Not only because of cash flow, but also due to car value getting diminished over time.
The priorities of stocked cars to get sold were blended into the products offered more actively. That helped to smarten the sales process and mitigate the stock aging problem.

Combining data and turning them into action
The described models and processes rely on data from multiple sources.
– ERP system (customers, products, orders, stock)
– PBX system (outgoing and incoming phone calls)
– metadata and model parameters

A data warehouse has been set up, which ensures data processing and cleansing on a daily basis. Interactive reports were created, which let the users see their list of customers to be contacted, as well as recommended cars to be offered.

Sales force team engagement
The sales team has been making phone calls to the current and prospective customers. They had to be trained and instructed in using a simple yet efficient dashboard. The dashboard lists their daily tasks and gives all team members a comparison how are they standing against each other.
Several team members used to be slightly reluctant to use a new piece of technology, but they saw good work results of their colleagues, everybody adopted the new way of working in the end.

Issues we faced
Perhaps the biggest issue was data quality. The list of customers and their contact phone numbers within the ERP system was very unclean. Duplicated customers, different phone number formats and matching them to the call duration records from the PBX system was challenging.

Despite the uneasy situation with incomplete and inaccurate data we were able to deliver the model and execute on it. Our data-cleansing procedures made a fair part of questionable records useful again.

ROI evaluation
When the sales force team previously used to work with customers as per their best judgement, they achieved average revenue per sales call about €13.

Then our Behavioral Prioritization. Product Recommendation and Stock Optimization models got introduced. Salespeople started to work based on scored customers and recommended times to contact them.

Despite a high inaccuracy of the contact list and considering only every second or third recommendation – since then the average revenue per sales call changed to approx. €21.60.

Your biggest loss is your unearned revenue. How much does every day of unused data potential cost your company?

As you can see, it is not rocket-science. You can create and apply a similar system in your company (probably in several months) on your own. Make sure that your data architect creates a modular, transparent and well-documented system able to include and blend multiple behavioral models. Also be ready to spend most time on data preparation, quality and workflow automation.
If you don’t feel like creating and maintaining such a system on your own, you can also hire an external data integrator.

Or contact Insightee today – to get your data crunched quickly, easily and with a full potential used from scratch.

Unlike other BI vendors (usually offering data integration and some fancy analytics) – Insightee is different. The “prescriptive” Insightee actions will directly turn your ERP data into extra revenue.

Dogfooding the Czech Parliament

Dogfooding the Czech Parliament

The Parliament of the Czech Republic, similarly to other democratic countries, is obliged to publish open data. Favorably for analysts, they do so on a very granular level. Individual votes of every parliament member are recorded and made available to the public for further analysis. This article describes how we processed the Parliament voting data and made its members engaged.

The initial idea of creating a public portal kuloary.cz analyzing and clearly presenting parliament members behavior was driven by our will to showcase how we can do analytics that differ and matter.

We decided to process the open data of the Czech Parliament on a daily basis and turn them into interactive dashboards revealing patterns coming from attendance and individual voting throughout the current election period.

Multiple Power BI dashboards backed by MS SQL Server data mart compare the members’ attendance rate, valid voting rate, party coherence rate and other. Figures can be analyzed globally or with regard to a specific political party.

From website visits we saw that people are mostly interested in the “attendance at work”. Which is a pretty basic metric, yet it seems to be very important for the general public.

Later on we met several Parliament members to collect feedback on Kuloary and to hear their perspective. The discussion has been very constructive because we could compare in detail the internal Parliament reporting with our findings. A rich backlog for further development emerged, but what was important, we validated the figures.

The most surprising moment however was when we learnt that Kuloary was being regularly used by many parliament members themselves as an extension to the existing, one could say basic, Parliament reporting. We learnt that politicians refer to Kuloary in their online campaigns, internal newsletters and even in TV. That’s been very satisfying for us as authors. And the regular website visit rate just confirms that.

A passionate analyst though doesn’t stop in such a moment and digs deeper. We looked for indications of impact on politicians knowing they were being observed more thoroughly than ever. One of the parameters we measure in time is the Valid Voting Rate. It is the number of cases when members validly voted YES or NO compared to the number of votings.

In other words – it seems that the parliament members became more reluctant to abstain from voting. It is hard to argue whether that is a coincidence or not. Nevertheless we could observe that since the release of Kuloary portal in July 2018 the Valid Voting Rate grew from approx 54% to 70% within 12 months.

Legally it has no impact whether a member votes NO/against or abstains from voting. But the NO also means that the politician made a decision and intentionally expressed his/her opinion. It is a sign of a”response-ability”, which is the true meaning of responsibility. Thus we can conclude that once the Czech parliament members realized that they were being analyzed, their behavior changed towards a higher responsibility.

Let the case be an inspiration for other democratic countries and a study of how a behavior exposure can eventually trigger responsibility.