Although every journey to becoming data-driven is different in the details, the fundamental problems and challenges are very similar.
After conducting over 30 projects for startups, we’ve identified 3 categories of challenges that innovative, newly established, and rapidly scaling companies face.
Let’s look at them
Startup Challenge I: User Acquisition.
The era of achieving commercial success with an app based on a good idea and product is long gone.
Because of major players like Apple, Google, and various ad networks that dominate and control the market, your chance of succeeding merely because you have a high-quality product is very low.
Add the fact that the lifecycle of a mobile app is often very short, which is especially visible in the mobile gaming market.
Despite high install numbers, most users only stay for a short time in the app. Significant user drops occur rapidly, with only a tiny percentage paying and remaining active over time.
|To stand out and prevail in the competitive mobile app market, startups must spend wisely on user acquisition and efficiently monetize those users in the initial hours and days and over the longer term.
And that means as a startup, you need to…
Measure your user groups by cohorts to see if they bring you lifetime value and ROI.
Micromanage every campaign and do AB Tests.
Build predictive models to maximize the campaign’s efficiency.
Understand whether your Revenue Over Advertising Spent (ROAS) is positive.
And you can’t do it without proper data and tech stack.
The solution: measure, test, and optimize.
In our approach to optimizing user acquisition and monetization, we start by pulling user activity data from the backend, mainly using Firebase.
This gives us a detailed look at every user action, which we blend with Google Analytics insights for better campaign targeting.
We also use data from Mobile Measurement Partners (MMPs) like Adjust or Tenjin, which are great for determining where the users are coming from and how effective UA campaigns are.
But the real magic happens when you put all this data together. Utilizing BI and Machine Learning tools like Big Query, we’re able to determine
- how long users stick around (retention),
- how much they’re worth over time (LTV),
- how our ad spending is paying off (ROAS).
Equipped with this knowledge, startups can build predictive models of how valuable a user will be just by looking at their first few days in the app.
Then, they can iterate paid campaigns accordingly and target people more likely to leave money in the app and become loyal users.
Finally, with data visualization tools like Google’s Looker Studio, management gets easy-to-read dashboards that show the actual numbers and the predictions to see how things are going and take necessary actions based on data.
Do you want to get a real-world glimpse of how such implementation works? Look at the Funcraft Case Study, where we show how using Google’s ecosystem tools helped the startup’s management effectively handle User Acquisition.
Startup Challenge II: Rapid Scaling.
The app is smooth and functional, but suddenly, it stops working – because, in a short time, the number of users increased by 5000%, for which the app’s architecture is not prepared.
This is bread and butter for startups, but also nothing unusual for regular companies that grow organically but, for some reason, suddenly experience an unexpected surge in customers.
Then the question arises…
|How do you develop an architecture that’s robust enough to handle millions of users without interrupting business operations while ensuring it’s scalable and prepared for future development?
This was precisely the challenge faced by one of our clients; a few hundred users used their gamification app daily.
But one day, a retail client brought in 6 million users in a single day, causing their existing Postgre database system to fail.
This sudden growth highlighted the need for a more robust infrastructure. We had to build an effective and scalable solution ASAP
The Solution: seamless migration and low-effort scalability
To address this without disrupting ongoing operations, we initiated an emergency migration to Cloud SQL, creating a seamless data replication in a new Postgre setup within the cloud.
We didn’t touch their production system, which allowed the startup’s operations to continue uninterrupted during the transition.
Next, we transferred the replicated data to BigQuery and set up Looker Studio dashboards. The two tools from the Google Ecosystem provided straightforward, cost-effective business intelligence solutions.
Within a week, startups can quickly get over even the most overloaded systems using nothing but a simple infrastructure copy-paste and a little bit of smart data engineering.
But you also want to make the architecture scalable to prevent such scenarios in the future. And without investing too much in the process, if possible.
In this specific case, we achieved this by implementing Cloud Functions for FTP uploads, facilitating easy data integration into their systems.
For another set of clients, we embedded Looker Studio dashboards directly into the gamification startup’s portal, which was visible and interactable but managed by the BI tool, eliminating the need for extensive programming for each new report or chart.
For a major banking client that operated within the Microsoft ecosystem, we integrated the capability to share reports using Power BI, meeting their specific ecosystem requirements.
Whether a startup or an enterprise, you want to balance customization and scalability, avoiding the need for overly complex tools like orchestration or terraforming yet effectively addressing the diverse needs of different clients.
Startup Challenge III: Building a cloud architecture from scratch
When companies merge, they often form a new subsidiary, which is not a classic startup but faces startup challenges of creating its own infrastructure, software, and processes from scratch.
|The big issue is integrating various data sources from the ‘old world’ systems of the merging companies, such as telecom billing, CRM, and network data systems, and then transitioning the entire setup to a cloud environment.
The challenge was even greater in this case than in typical startups.
First, the complexity arose from the technological integration and the scale and diversity of the legacy systems involved.
Second, post-merger, there’s often hesitation to expand the workforce, and in this case, the parent companies were reluctant to hire a new data team.
Thus, the startup team had to navigate the introduction of cutting-edge cloud technology with limited resources.
They sought external expertise to build this new infrastructure.
The solution: On-demand data team as a service.
In this case, we moved the legacy systems to a cloud environment, employing various tools beyond standard cloud functions.
We used CloudArm, featuring a secret manager to address stringent security requirements, along with a data form, excluding DBT.
Corporate procurement and approval norms influenced the decision-making process for Microsoft cloud services.
Large corporations often find it easier to approve technically less optimal solutions but in line with the corporate ecosystem.
For instance, the company’s existing alignment with Microsoft products like Office 365 and SharePoint influenced its choice of Power BI, aligning with its familiar operational environment.
One way or another, the business outcomes were substantial.
Within a few months, we established a comprehensive data platform, integrating ERP and network systems without needing additional internal resources except for a project manager.
This rapid deployment starkly contrasted the time-consuming process of building an in-house team, which would have been redundant after project completion.
Post-implementation, the focus shifted to building reports and analyses and training the business team to utilize the new platform, exemplifying a successful integration of technology and corporate strategy in a complex, multi-dimensional environment.