Growth Leaders Forum: roundup on Analytics (and decision making)
Last week I organised a virtual roundtable with a number of fellow experienced Growth professionals with the idea of sharing thought ideas and learnings. The topic under discussion was Analytics.
Over time I have noticed that very often the starting point of my projects is analytics and that in most cases this area is a big challenge for organisations. I remember being part of another roundtable years ago where at the end I asked everyone how would they score out of 10 the analytics capability of their company: the scores went from 2 to 6, (with just on 6).
This was a great opportunity to ask the same question to a group of people whom had all worked for many different firms – the outcome was similar. Most firms’ analytics solutions are not fit for purpose.
Personally I find this staggering. Keep in mind that between the five participants of the calls we have seen over 50 companies, covering most sectors and of different sizes. From early stage companies to established tech enterprises. We all agreed that adequate analytics was a basic requirements for company’s survival. After all without data you can’t learn, so you can’t improve, you can’t react and fundamentally one can’t make the necessary decisions to run a business, let alone grow it.
Gianluca Binelli, ex-Google who now runs a performance marketing agency, has over the past few years worked with many different companies and he confirmed that step one for any project was auditing analytics. He also shared how rarely the solution in place is in good shape.
So what is a satisfactory solution? What is a 6 out of 10?
Gianluca said that this often is just about getting the basics in place really well and tying marketing data to finance data. Having an accurate idea of marketing costs, cost per acquisition and lifetime value is a great first step – but that isn’t so easy. Different data sets need to be tied together and be in a format that is understandable and accessible quickly and regularly by the relevant people. The skills needed to put this in place are a combination of maths, marketing, technology, accounting and general business understanding.
Oren Greenberg, who has worked as a high level consultant for the past 8 years, shared with us his experience when working in a high growth FinTech: the analytics in place were some of the best he had seen. That was because there was a person dedicated to the analytics project from early on in the company and the individual has a master’s degree in mathematics and a good understanding of business.
At this point we agreed that generally the best analytics were found in mid-sized tech companies. These are organisations were there are sufficient resources that can be invested in solving the data problem, a relatively simple business model, (one product/business unit), and not too many legacy processes in place.
At this point Emilija Frew, with whom I had worked with at HSBC where she acted as Global Head of Digital Commercial Performance, raised a killer question: say that firms did have perfect data and information flow – would the right decision be made?
This brought up two other key points:
- Data interpretation
- Company culture
Oren shared with us quite an interesting data point: 70% of academics don’t fully understand statistics.
In my experience data has this strange property that the more sophisticated you get with it, the less it becomes certain and absolute. It becomes less black and white and more open to interpretation and that is where an understanding of statistics becomes important.
Of course when data becomes open to interpretation it becomes easy to have a conflict of interest and read results in the most favourable possible way. There are two ways to mitigate this: culture and designing a good analytics operating model.
Here enters the North Star Metric, (NSM), a smart way of aligning long term growth, processes and culture. The North Star Metric is the single metric a whole organisation works towards – and it’s not revenue nor profit! The NSM concept that when correctly applied can be extremely powerful and it focuses on value creation rather than the consequences of it. Revenue and profit are consequences of value being delivered to customers. If the firm’s mission is to help people find a taxi, then rides delivered is a good metric. Youtube famously used numbers of video minutes watched. Basically the NSM needs to be aligned with the company’s mission.
Around the NSM the analytics team/person should build a Growth model. This is a set of 4-6 metrics showing the KPIs around the value creation flow of the company (the pirate metrics AARRR are on common implementation of this). The KPIs might be more if the business is a platform or marketplace. These represent the cockpit of the company and are very valuable in giving a quick health check on overall performance, (here is a more detailed piece on creating a growth model.
The Growth model, when properly built, allows the team to run scenarios and understand the impact of different activities on the NSM over a given period of time – this in turn informs prioritisation. As work and processes are geared to support the Mission, a strong and positive culture comes to live. This will ensure that once people have the right data they take action it in the best possible way.
Personally I think this approach to analytics that has a bias for action and a positive impact on culture, is a trick many corporates are missing out on.
Yara Paoli ex VP of Growth at Skyscanner shared the details about how they achieved that during her tenure at the travel metasearch. They embraced the methodology fully, aligning all KPIs, actually OKRs, to the value creation process aka the NSM, (not to revenue, revenue is a consequence of it), and designing a data empowerment solution for growth that allowed for local decision making (based on impact on the NSM and full understanding of AARRR metrics). Finally they created internal purposefully designed courses to increase the overall level of literacy around data analysis and growth.
This major data enablement process took two years and significant investment in terms of leadership focus, teamwork and enablement activities, but it proved pivotal for Skyscanner’s Growth.
When I asked her how she would’ve rated their growth analytics solution out of 10 she said “7.5 at best”.
During my tenure at IG I managed the roadmap for marketing attribution, working very closely with a team of data scientists. We managed to have good data round multi device attribution, multi touch attribution and life time value. We also had purposefully designed data dashboards for improved visualisation and decision making – yet some parts of the data accuracy could’ve been higher (we were at best 8/10).
We managed to have a good analytics operating model.
I previously wrote a blogpost about this here. The fundamental process for putting that in place is the following:
1. Map all decision makers
2. Understand what decisions they need to take
3. Agree on what frequency those decisions need to be taken
4. Contextualise the decisions to the overall business
5. Define the best set of metrics to report on
6. Understand what is the minimum level of accuracy required
We did not manage to have this model run through the whole funnel, but implementing this at the top of the funnel allowed us to scale activity by almost 300% on a global scale. After all data leads to learnings, which leads to improvement, (in a global organisation this should be coupled with decentralised decision making and knowledge sharing).
The benefit of the Analytics Operating Model is that decision makers are presented only with the metrics they need, thus simplifying the interpretation of the data and improving decision making.
Finally Emilija rightly highlighted that where possible the whole process can be simplified by taking company politics out of the picture and allowing machine learning to automatically optimise the delivery to customers.
We concluded our session talking about how linking marketing analytics and finance allows for a clearer link between actions and overall top line performance – thus enhancing actionability. However collaboration with finance often proved tricky for two reasons, first because marketers tend to have a low literacy level of accounting, (we need to improve here), and second because of the different type of personality types across the two departments; on more focused on growth, the second more on stability.
To any firm out there struggling with analytics I would say not to worry, almost everyone has those issues, but solving them is critical and while significant investment may be necessary, the prize is well worth it.