Robin Aro, CEO, Head of services, DB Pro Services Oy.
The morning starts as usual. I’ve got a long to-do list for the day, and the plan is to make progress on the new reporting framework, process data for analytics, and perhaps finally get to grips with that AI solution we’ve been discussing with the business for some time now.
Then a message appears in Teams: “The figures in the report haven’t been updated. Is there something wrong with the data?”
At that point, the plan for the day changes. First, we check the report, then the data downloads, and it soon becomes clear that one run has failed during the night. After that, the investigation begins. Was the problem in the source system, the interface, the data download, the database, or somewhere else entirely?
Meanwhile, other work is piling up. Progress on reporting grinds to a halt, the next phase of analytics is postponed, and the AI trial is once again put on hold. By the time the problem is finally resolved, the next issue is already waiting on the table. A new version of the software needs to be installed, a minor change is coming to one of the interfaces, access rights need to be reviewed, and the database’s performance has been gradually deteriorating.
None of these tasks on their own is necessarily a big one, but taken together they take up a surprisingly large proportion of one’s time.
When maintenance starts to take time away from development
With a data platform, the biggest challenge isn’t always the technology. Often, it’s time.
Data needs to be sourced from various sources, combined, processed and modelled. Reporting needs to be developed in line with business needs, analytics needs to be taken forward, and at the same time, time should also be found for new use cases for artificial intelligence.
At the same time, the existing platform and databases should simply continue to function.
However, it is precisely this ‘mere functioning’ that requires constant attention. Data loads must be monitored, databases maintained, changes to integrations addressed, and reports updated whenever the source system or business logic changes.
Often, the same small team is responsible for both maintenance and development. When a problem arises in production, development comes to a standstill. When development is pressing, maintenance tasks are easily put on hold.
Ultimately, both are competing for the same time, which leads to things slowing down.
Does the team need one more person?
At this point, one solution readily springs to mind: let’s hire a new expert.
A new member of staff could provide additional capacity and reduce reliance on individual key personnel. In practice, however, the question quickly arises as to whether there is enough maintenance work to keep one new member of staff busy full-time.
The requirements for maintaining a data platform vary. One day you need database expertise, the next data engineering, and the day after that knowledge of reporting or cloud environments. Sometimes everything runs smoothly for weeks without any major issues, and then suddenly you urgently need expertise that isn’t available within your own team.
A new recruit would therefore need to be skilled in many areas, but some of their capacity might still go unused.
Recruitment is not just about salary costs either. It involves recruitment, induction, training and the ongoing maintenance of skills. Furthermore, critical skills may once again be concentrated in a single person.
In that case, it’s worth asking the question a little differently: do we need a new person, or do we need the assurance that the right skills will be available when we need them?
Who responds when something happens?
When it comes to maintaining a data platform, one of the most important things is, ultimately, very simple. It must be clear who takes charge when something happens.
When a morning run fails, someone needs to spot it and find out why. When an interface changes, the implications need to be assessed. When database performance deteriorates, the problem should not only be noticed when users start complaining about slow reports.
The same applies to smaller matters. Updates, access rights and changes to reporting should not have to be remembered whilst juggling all other work, nor should they constantly be pushed to the bottom of the development queue.
At this stage, the ongoing service model is starting to look like an attractive alternative to expanding our own team.
The idea behind Data Platform Care is that the expertise required for the platform’s ongoing operation is available without the need to build everything in-house. This expertise may relate to databases, data flows, reporting, analytics, integrations or, at a later stage, artificial intelligence solutions as well.
From the customer’s point of view, what matters is not who carries out a particular technical task. What matters is that the matter is dealt with and that responsibility is clear.
More time for what the platform was built for
The aim of outsourcing data platform maintenance is not to take away the customer’s control over their own data or environment. The aim is to free up time for the work in which the organisation’s expertise creates the most value.
If an in-house expert is familiar with the business, the customers and the company’s data, their time is often better spent developing new analytics than analysing logs from a failed overnight run.
The same applies to reporting. If there are several development needs that are important to the business waiting in the queue, it is worth spending time resolving them rather than allowing day-to-day issues with the technical environment to constantly interrupt your work.
Artificial intelligence adds yet another layer to this. New solutions cannot emerge if all available capacity is taken up by maintaining the current environment. Furthermore, AI solutions require a reliable data foundation, just as reporting and analytics do.
Cost-effectiveness without excess capacity
An ongoing service can also be a cost-effective alternative to expanding your own team.
There is no need to recruit all the necessary expertise into your own organisation just to be on the safe side. The service can be scaled to meet actual needs, meaning the client does not have to pay continuously for capacity that may not necessarily be used.
At the same time, it is not just the expertise of a single new employee that is available, but, where necessary, a wider pool of experts from different fields. This reduces the costs associated with recruitment and induction, whilst also minimising the risk of critical expertise being dependent on a single person.
So it’s not just a question of whether to carry out maintenance in-house or outsource it. The more important question is: what is the most sensible way to ensure that the platform works properly and that your own team can devote their time to the matters that are most important to the business?
Peace of mind is also part of the service
When discussing the maintenance of a data platform, the conversation often centres on hours, costs and technical tasks. One benefit is then easily overlooked: peace of mind.
When responsibilities are clear, there’s no need to constantly worry about whether the runs will work tomorrow morning, whether a problem will be spotted before the report users do, or who will step in when a key member of your own team is on holiday.
A well-functioning data platform is virtually unnoticeable to the user. Data is updated in a timely manner, reports work as they should, the database remains high-performing, and new solutions can be developed without every production issue bringing everything else to a standstill.
That way, you know that a familiar and trusted party is replying to the Teams message, and there’s no need to interrupt your own important tasks. Instead, you can spend your time on the very question for which the whole platform was built: what new things can be done with this data?
That is what Data Platform Care is all about. We take care of the platform so that you can focus on making the most of it.
