MTN today released it’s results, and as always I read these with immense interest, and with a lot of passion, this is ultimately where my passion for mobile was born, so the affinity to the company I suppose will always remain part of my DNA.
Without going into financial analysis and extracting intelligence from the results and the accompanying narrative, I wanted to expand today on what some may deem to be a boring subject, that of knowing our numbers, and no I am not talking about the financial numbers here, I am talking about the numbers that result in financial results, the numbers that determine if we know our business or not, the underlying drivers of the business, that in most cases are ignored or invisible as we get involved in what we deem to be more important, when it is not.
My engines get all fired up on a number of topics, these being strategy and executing such strategy, the importance that people add to achieving results, and then the subject of this conversation, looking at the numbers and extracting insights and intel from such.
I am still amazed at how often CEO’s and Executives do not know the underlying events in their operations, especially relating to customer and channel behaviours, and then at month end tries to explain the financial results based on believes or assumptions, with the most favourite one being “the economy” and the “regulatory challenges” when things go bad or below budget.
So within MTN results there is reference to an “IGNITE” program, basically the program aims to really drive operational and financial efficiencies, which is vital in todays business environment, and all operations and business should have a similar program that relooks the business line-by-line on the Income Statement, the Cash Flow and Balance sheet, with aggressive targets, as stated in the MTN results as well.
One specific line however caught my attention, and it read as follows:
“use more advanced data analytics to better inform our decision making, particularly around customers and network deployment”
THIS is where my engine really get’s started up, as so many times have I seen that decisions are made on believes and field marketing studies, without looking in detail at what the actual numbers say, and when numbers are extracted it is extracted at the point of time, and in most cases it is made and/or presented fit for purpose, a single snapshot in time, a few tables and excel sheets and that is it, decision made, programs implemented an at month end or even later we will have good or bad results, if it is bad we go back to economy and/or regulatory fallback in the boardroom presentations.
I believe that each C-Level in each operation needs to have an in depth view of what is happening in their business, and to really know their business, lets see if you look at the things the way I look at it and if you can put the checkmarks to the boxes below, as I expand on what I constantly look at, how I look at it and why I look at it, and then what I do with it…
Now from here you will read a lot of “I” in the different sections, this is only for article purposes, as there is a great team behind these that makes all of this possible, so read the “I” in this as a collective rather than an individual element, my people are vital and very important to me, as stated previously it is another one of those points that get me all fired up, in fact I believe people should be a value on a balance sheet, but that is the subject for another day 🙂
My starting point on any of the dashboards and management tools I created starts with the obvious, that of the number of customers on the network, but I am not interested in the Revenue Generating Subscribers (RGS) numbers as reported by the industry, by the time you see negative movement on 30/60/90 days your problem is already old, I look at only the VLR attached numbers, by month, by day and by hour, and I have my dashboards set up so that I can see the trend go back as far as possible, my current dashboard goes back 5 years, and I overlay every marketing, competitor and other major events over this dashboard, so that I can analyse what impacted any customer movement, I also use it to analyse a number of other trends of different programs, strategies etc, even the weather (rainy season as an example), as well as downtimes that occurred on the network, and if I see any movement the operation is set up to react and adjust swiftly.
Similarly to above I look at the recharges per month, day and by hour, yet again to determine the customers behaviour across the network as an average, but I drill down deeper as I have learned that averages can hide a lot of vital statistics, thus I created a segmentation model, that splits the VLR active base in exact 10 parts, from the highest segment to the lowest segment in terms of recharge value, each representing exactly 10% of the base in terms of customer numbers, I replicate this segment model in terms of usage as well, and I know exactly what customers are spending their money on, by segment, when they are recharging and when they are using these recharges, On Net, Off Net, International, Data, SMS, Value Added Services by the service utilised, and what the average value in wallets are by segment.
I use the above analysis for a number of areas in the business, from determining the number of days of stock in the market, this by overlaying the channel data that I have in terms of reseller wallet values across the different values, the datasets I designed can drill down to a per region or per site basis, I know exactly how many active points of recharge I have in each area, the volume of transactions they are making and similar to the customer model i have segmented these overall and by area into the segment model, thus if I need to drill down I can see what happens in each site on the network from a commercial perspective, I also use this data to determine if we have an oversupply or undersupply of recharge point sin each area, as well as what services may require a bit more of an education and/or marketing push.
Also within the above segment model I can focus on the actual segments in terms of usage factors of each services, and I can drill down in a further 10% decile split of the segments, vital info for loyalty or bonus programs etc, but even more vital when I do the margin analysis of the customer as the model allows for overlaying of each cost item on the various revenue streams, such as dealer discount, regulatory fees, interconnect and others, thus I can keep a constant eye on gross margin by segment and sub segment, the same goes for sites in determining not just the relevant contribution to revenues, but also to margins overall, this prior to and after OPEX items, both equally important.
The model I created also looks at the details of usage, as an example I can tell you on International not just what destinations are called, but what day and hours they are called, how many distinct MSISDNS call these destinations monthly, how many are repeat MSISDNS, number of calls they do on average by segment, the duration of these calls and the spend and margin averages, and I have discovered gems when putting together the international product structure, and avoided disaster when changes from carriers are implemented in terms of the underlying cost of the traffic, I have the same data for incoming international traffic, and I constantly analyse the trends, as I have learned that the smallest change can have a dramatic effect in terms of revenues and/or margins.
It is also at this point worthwhile to note that I believe that most networks today needs to invest urgently into big data technology, as the number of data points that exists are vast, and believe me when I say the smallest change you make on pricing or product structures are bound to impact another parameter elsewhere, it is virtually impossible to monitor each point in a network, but I have seen many cases of slight changes having an impact elsewhere that were invisible to managers not used to looking at each and every trend, and when investing in this big data technology it is equally important to invest in AI technology as well to “learn” the data all the time, analyse movements and present the recommended actions, and ultimately make adjustments on the fly, this is the future of dynamic customer/tariff/product management in telecoms, invest today and reap significant benefits going forward.
Besides the obvious parameters such as voice traffic I also analyse the recharge and usage behaviour of data customers, once again by 10% deciles segment, I know what usage is spent on by application (such as whatsapp), durations, average spend, average margin, time of day, day of week all the way through to the cumulative monthly and yearly analysis.
From a device perspective I know exactly what the spend is for each type of handset, smartphone, make, model, how the brands play a role in the value of the customer, the tenure of handsets by type, 3G penetration by site in terms of handsets, active and non active 3G users, and the associated margins with each, vital statistics as an example when you are trying to determine that next handset promotion, I even track what happens to “hand downs” (customer old handset) when customers change handsets, and can tell you if the handset remains within the customer group or not, thus the customer added possibly additional value that is not always evident when looking at normal data.
I also analyse new customers coming onto the network within above handset parameters and can determine the breakdown of customers joining by handset type model etc, overlaid with the average handset selling price in market, it gets very interesting when you analyse the average revenue and margin as a percentage of these handsets as an example, as well as the time it takes customers with data handsets to make the first data connection, as well as what that connection is, insightful, powerful, vital.
From a mobile advertising perspective the data analysis becomes even more critical, it is more value adding if you can provide would be advertiser with detailed on information beyond demographics, and unfortunately I would absolutely love to expand on this point but we are in a competitive environment, and unfortunately cannot expand today, but I will share this in the future.
Most of the above information I also use with my channel partners when we structure their contracts and incentives, every single clause in agreement is somehow tied to a number that can be measured, against an objective that needs to be achieved on the Income Statement, Cash Flow and Balance sheet, if you do not have such an agreement then the strategy and deliverables will always be a guessing game, that will ultimately put strain on the relationship and never get the results, and you would have to fall back on economy and regulatory challenges at month end.
I added elements such as Cost per Hour (CPH), downtime, congestion, Erlangs, RNC to the different overlays, to ensure that we do not miss a beat, including average tariff and average revenues margins per erlang, by segment, region and site, amongst others.
From a care perspective even the call centre numbers are analysed, down to a lot of detail beyond the traditional call centre reports, yet again by segment, type of device, query and monitoring if customer remains or falls of network once call was made to call centre, a better indicator than most, not that I do not believe in concept such as “Net Promoter Score” (NPS), but I prefer hard data.
Looking at MTN again and services such as Mobile Money, MTN Games Club, Mokash, Remittances, Jumia, Snap and others, I would absolutely love to get my hands on this data and analyse the value that each adds to the underling customer, usage patterns, behaviours and a long list of other elements that come to mind, it would make for extremely valuable information indeed.
I would love to add all the parameters I use, but the post will be too long and as said, some of the points are better kept close to heart 🙂
Wish that each of you discover your ultimate dataset, bottom line is know your numbers, and know them well.