Expect the Unexpected — How to Introduce Data Management in a 600+ Scale-Up Company

A few years ago, I wanted to find out the number of Netguru employees that had no billable assignments or were available for new ones, with the aim of improving our scheduling process. At that time, our internal database was unable to present these insights in a single place. Since it was critical for our performance, we had to either improve our current system or create a new one from scratch. Eventually, I decided to work it out by developing our system into a company-wide command center, where a new ‘internal’ assignment for all the availability/non-billability cases I was looking for was just one of the features.

It seemed to work fine until an internal workshop we had a few weeks ago. In the middle of the discussion we realized that even though we (probably) had all the data needed, we were unable to drive some insights we were looking for. And the reason was the solution I came up with two years earlier. So we had to rethink our approach once again. Every scaling-up company faces similar issues — here’s our story and how we’ve managed to tackle our problems.

In the past, companies used to operate in a highly predictable environment, where projects and contracts were signed for years ahead, providing stability and security. Under such conditions, both revenue and costs used to be highly predictable.

Those days are long gone, as the world today is way more dynamic than ever before. From a business point of view, this dynamicity comes with two crucial elements:

• Fast-paced environment. In order to survive, modern companies need to experiment and iterate; such an environment requires switching to short-term operations (months instead of years) and fast idea validation (MVP) for healthy cost and profit management.
• Flexibility. Digital transformation made it possible to react to market trends and tendencies literally within minutes. This ability makes it easier for businesses to tap into new, time-sensitive opportunities, and to minimize losses in case of business risks and failures.

Netguru is a digital consultancy, offering services ranging from product ideation and design to development and maintenance. Long-term cost predictability is one of the key factors for ensuring stable growth in technology-driven businesses like ours. The more accurate our predictions, the higher the chance that we deliver business KPIs, hire new talents, or move resources internally where they would be needed the most.

The problem is that to make accurate process predictions, you need to continuously look into the future and reduce the degree of uncertainty. There are two basic means that can help you here:

• Building predictions on accurate and qualitative data;
• Shortening the recruitment life cycle.

The above is true for every business I can think of. Let’s take UBER for example:

• To provide a reliable time of arrival, the UBER app needs to analyze real-time traffic data imported from multiple sources;
• To cover a rapidly growing demand for rides in particular areas, UBER needs to adjust the recruitment process and diversify sources of cars and drivers;
• To increase the NPS, improving the UX and generally speeding up the whole ride-ordering process might be crucial.

Photo: Achim Pock / unsplash.com

Combining the three things above offers the best way of preparing your business for the future — you can go from losing a significant client to embracing new, potentially profitable solutions by existing talents. Of course, sometimes it’s hard to be this flexible, especially when:

• Saying ‘no’ to a client could result in losing their trust;
• We cannot afford to lose our margin.

Apart from these situations, the maximum ability to adjust to the current situation is crucial for ensuring growth in modern companies. It applies especially to businesses offering digital services — like Netguru — where supply and demand cycles are much longer than the predictability of winning and maintaining a certain project or client.

So, how do we tackle such a challenge at Netguru? There are two foundations that help us not only secure growth but also to grow faster than other market players (120% growth in 2018, 50% growth target in 2019):

• Agile. Thanks to this company-wide methodology, we’re able to leverage the full potential of our talents, e.g. supporting our development consultants in embracing new technologies that are/will be in demand. It also makes it easier to find weak spots on each project and choose the right tools to solve any resulting issues.
• Data management strategy. I’ll focus on this point from now on.

At the beginning, we used to operate on multiple apps and systems. There was no unified environment in which data would be collected, analyzed, and transformed into insights. As a result, there was no single point of truth, and every tool could lead to different insights than others, causing information chaos and impacting our prediction accuracy.

The first step to change it was to actually realize that we already have the data we need.

What was blocking us was the lack of a data management strategy. We needed to start thinking about data, to make it the starting point of all our operations, but at the same time, it was important for us that we cut down the number of tools to an absolute minimum. As a result, we’ve built an integrated data management ecosystem that allows us to collect and analyze the right information, aggregating data from each business unit.

Here are a few examples of how it worked before and after we introduced the new data approach to each business unit.

1. Headcount planning

Before: we used to recruit talents according to our current needs and to what we felt could happen in the near future. It was all wishful thinking at its best.

Now: we can predict the talent pool we need to hire before every quarter kicks off, broken down into technology stacks, services, seniority level, etc. Thanks to improving the predictability of our headcount, we can secure the proper size of each team servicing our clients. In case that any events occur during each quarter, we’re able to adjust the headcount plans thanks to the Agile methodology. We use Workable for recruiting purposes.

2. Inbound lead generation

Before: we used to focus on getting more and more leads and trying to make them interested in our offer.

Now: we’ve built a complex inbound marketing strategy that automatically picks out high-quality leads, and carefully nurtures them along the marketing funnel until they’re ready to become a sales opportunity. There are multiple content campaigns we use (from Employer Branding to React Native Development) — all built and managed using HubSpot.

3. Predicting project success rates

Before: we were only able to predict how a project would fare with a 40-percent accuracy. We used top-quality tools, (Salesforce, Jira, Github), yet they didn’t converge well. The outcome was that they wouldn’t give us the data we needed to create metrics and make good decisions.

Now: since Salesforce was already in use at Netguru, we moved to this tool with scheduling and integrated other data points (Jira, Github) with it. Displaying the data in Salesforce in the right context (business and individual project) improved our processes, helped us organize our work, and had a very positive impact on our predictability. As a result, over a few months, we bumped our project success prediction accuracy to about 81% (2018).

As I mentioned earlier, it’s important to collect only the data that helps you make smart decisions. It might be tempting to collect all the information out there because you’d never know when something might become useful. This is a mistake many businesses make, which renders them unable to drive strategic insights. You can have a set of 100 KPIs that provide no insight and 10 that really help you grow. That’s why we’re constantly struggling to make our system clear and simple, removing all the data sources that appear unused or unnecessary.

For example, if you want to define your company’s customer satisfaction level, there might be a lot of different factors you could look at. In the past, our way was to track many of these, including talking to our PMs or guessing if a certain client is satisfied based on their tone of communication, etc. It usually led us to a broad set of “really important data” and no actionable knowledge at all.

Today, all we do to measure customer satisfaction is a quarterly NPS survey, which has replaced all the other KPIs. Whenever we score less than 9 from a customer, we work on improving their satisfaction for a whole quarter and then ask them again to see if there’s any improvement. At the end of Q1 2019, our overall NPS score was 68.5 — way above the industry average (24 for IT services, 61 for the tech industry). It basically means our clients are more than satisfied, which we would never really get to know without the NPS.

To sum it up, by introducing a data management strategy, we managed to improve nearly every strategy process at Netguru, including:

• increasing our ‘bench’ predictability for months ahead;
• improving recruitment capabilities, from defining the talent pool we would need in the near future to the right balance of staff seniority;
• improving the accuracy of our regular cash flow analysis as a part of ensuring healthy financial growth;
• and many more.

Being able to “see” so far into our future gives us an unmatched comfort in terms of planning our work and our future assignments. After the first 9 months of operating within this strategy, we managed to increase the predictability of projects by 102 percent. It had a direct impact on our growth and helped us stay on the market lead position. Eventually, it was also appreciated by our clients because we were able to provide high-quality services within the expected timeframe.

Obviously, it will never be a job complete. Constant improvement and verification are a must for this system to work properly. Otherwise, we’d find ourselves with a bunch of inaccurate predictions, not being able to react before it’s too late.

Some of the pitfalls you should expect and get ready for:

1. Data is always as good as the processes and people behind it. Stay aware that the information you see will never fully represent reality.
2. The truth behind the data changes over time. For example, lead qualification might become more strict, which means you’d see fewer leads, but each of them would be of a higher quality. The same goes for pipeline value and price increase.
3. High-growth companies like Netguru change fast. It is hard to use data to compare YoY numbers, as you’re basically comparing different companies.

If you’re interested in reading more about how we improved our project management predictability between 2017 and 2018, check out this case study posted on our blog.

And good luck!

***

We’re not going to stop here. Our next steps at Netguru are:

• Developing specialized analytic competence;
• Systematically simplifying the whole ecosystem to make its management easier and cost-efficient;
• Applying Machine Learning algorithms to the collected data in order to improve all the above.