Editor’s note: Irene provides a comprehensive overview of data analytics outsourcing, the advantages it brings, and the concerns it raises. If you consider outsourcing your data analytics solution, feel free to explore our offering in managed analytics services.
Own-or-rent decisions are well applicable to the analytics domain: businesses need to choose whether to opt for building their in-house data analytics solutions or outsource the service. As the global market for the latter is expected to grow annually by 22%, I suggest you have a closer look at data analytics outsourcing.
Data analytics outsourcing is the cooperation model under which a company entrusts a service provider with its data and gets access to insightful reporting. At the same time, the provider takes care of everything else: infrastructure setup and support, data management and data analysis.
Two examples to envisage the service
Imagine an FMCG manufacturer suffering from unstable sales performance but unable to either run diagnostic analytics to identify what triggers poor performance or make accurate sales forecasts. At the moment, the manufacturer has neither time nor budget to grow their in-house data analytics team, so they turn to analytics outsourcing, hoping for fast insights.
To start delivering the service, the vendor needs the manufacturer’s data. They clean and organize it, and in the 4-week time, they provide the manufacturer with access to the Power BI reporting tool with comprehensive analysis revealing the reasons for sales instability and the potential to increase the manufacturer's sales by 15%. The vendor also provides the sales forecasts powered with data science.
Image credit: Microsoft
Another example demonstrates big data outsourcing. Think of an online entertainment provider having hundreds of thousands of visitors daily who search for the songs, the singers and the playlists they like. Meanwhile, the provider collects tons of data about their visitors – the day and time they visited, searches they made, singers they listened to, songs they liked and disliked, playlists they created, and a plethora of other data.
The provider runs only a basic analysis of this data. For example, they calculate daily traffic, define the busiest time when the majority of users turn to the service, identify the most popular songs and singers. However, the provider finds these insights insufficient for their business needs – they want to have their customers segmented, understand the preferences of each segment, stay updated with their behavior and personalize the service as much as possible. To achieve that, the provider’s current big data solution should be enhanced with big data analytics capabilities. As the solution reworking will require months and significant investments, the provider decides to outsource their big data, pay a subscription fee, and start to get the required insights within a couple of weeks.
Let me make it clear, to benefit from fast insights with data analytics outsourcing, you should be ready for long-term cooperation (typically, 2+ years) and a monthly subscription fee. A subscription fee covers data preparation and management activities, as well as the agreed number of regular and ad hoc reports.
And what concerns an alternative approach to data analytics, an in-house solution would require the costs of the solution’s design, implementation and support, the costs of hardware or cloud subscriptions and software licenses, and the costs of keeping an in-house analytical team.
A ready-to-run service
Contract signing and service deployment by a vendor, which usually takes from 6 to 8 weeks, is the only thing that keeps you from getting access to the batch of agreed reports and getting the value out of your data. For comparison, the design and implementation of an in-house BI solution that would enable the same reporting can take from 6 to 8 months.
Industry-specific best practices
A professional outsourcing partner elicits your requirements for reporting and brings in industry-specific best practices. For example, at ScienceSoft, when working with clients, we provide detailed consultations on what kind of reports will bring more insights, what data sources should be used to create such insightful reports and much more.
Technology and process expertise
Your outsourcing partner saves you the trouble of exploring, say, the differences between Apache Cassandra and HDFS (two technologies that can be used to build a big data storage) by taking complete care of the technology side. The vendor decides what technologies to choose to process the data you have and with the performance you require. They have hands-on experience of implementing, integrating and managing different business intelligence and big data technologies.
For example, one of ScienceSoft’s clients had clear requirements regarding the analysis but wanted us to take the responsibility for its technical embodiment. Thus, a US management consultancy could get pre-built reports with valuable insights about their particular aspect of interest skipping all the technicalities.
Here, let me address the questions that companies most commonly ask before opting for outsourced analysis.
How to choose the right outsourcing partner?
The first thing I suggest doing is to check the vendor’s expertise. For this purpose, years of experience in data analytics, as well as the portfolio of implemented projects will serve the best. While studying the portfolio, pay special attention to the projects that belong to your industry – it’s the simplest way to understand whether the vendor will be able to bring valuable insights to your business. It will also be wise to check the vendor’s partnerships and certificates.
Another important aspect is the quality of collaboration – check such basic things as speaking the same language (literally) and the ability to communicate within your working hours, as well as more advanced things like friendly and proactive approach and effective conflict resolution.
Will our data be secure?
Many of the companies, who commission ScienceSoft as a data outsourcing partner, feel uneasy about sharing their data, which is absolutely normal. Being fully aware of the importance of this issue, we ensure the same or even higher level of data security than when the data is stored internally. So, my advice to you is to cover the aspect of data security in the outsourcing contract: clearly describe the requirements to the environment where the vendor should store your data, as well as security measures and the liability imposed on the vendor with regard to ensuring data security.
Should we be actively involved in the process?
Normally, a company’s active involvement is required at the discovery stage, when the outsourcing partner scans the as-is situation to understand the analytics needs of the business better. Based on the findings of the discovery stage, the partner prepares a service-level agreement (SLA) and agrees upon its terms with you.
A good outsourcing vendor self-manages their work due to mature and transparent processes. They take over the full responsibility for the service delivery and its quality. However, I recommend that you establish continuous communication with the vendor, for example, provide them with timely feedback on the new features or evaluate the vendor based on the criteria defined in the SLA.
Will we get the description of the analytical models used?
The vendor doesn’t have to share with their customers what analytical models they tried and which ones were recognized the best. The same applies to the architecture of the data analytics solution or a deep neural network, the hyperparameters and configurations. In a word, the vendor has the right to leave all the technicalities behind the scenes (unless otherwise specified in the agreement).
Below, I’ll briefly describe the essential aspects that should be covered in a contract with your outsourcing vendor.
Key performance indicators and service level objectives
With wisely chosen KPIs and well-defined SLOs (service level objectives used to evaluate the vendor’s performance), you can clearly express your objectives and set the direction for your outsourcing partner, as well as exercise control over the service rendered. In their turn, the outsourcing partner knows what you expect from them, and they can organize their work and allocate their resources in the way they find most effective for reaching the objectives.
It’s important to provide each SLO with a thorough description and a measurement interval. Have a look at some examples of KPIs and SLOs:
Example of an SLO
An ad hoc report of medium complexity will be provided within 2 working days from the date of request.
Over a month
The web interface (Microsoft Power BI) will be available to end users at least 99.9% of the time and provide the correct information in full.
Over a quarter
Timeliness of service delivery
Regular weekly reports will be updated each Monday at 8 AM.
Over a month
Reports and communication
In the SLA, specify how frequently the outsourcing partner should provide you with a report on their work. I also advise you to particularize how often your communication with the outsourcing partner should take place, what project roles on your and the partner’s side should communicate directly, and what should be the aspects of their collaboration. For example, the users in your business departments can address the vendor’s analysts directly to provide their feedback.
The opportunity to cancel the contract for non-compliance
Make sure that your outsourcing contract contains a clause stating that your company can terminate the contract if the partner fails to meet the deadlines or other SLOs. There’s another reason for paying due attention to the detailed descriptions of SLOs, as well as setting forth the agreements on reports and communication.
So, does your business need data analytics outsourcing?
To answer that question, have a look at some strong indicators defined by ScienceSoft’s data analytics team:
- You want to get value out of your data fast (within weeks, not months).
- You don’t have time or resources to develop and support an in-house solution.
- Data analytics competence at your company is yet to be developed.
If you have found yourself in any or all of the above situations, my advice to you is to consider data analytics outsourcing, so feel free to contact me for further details.
Get a ready-to-run service from a vendor with 30 years of experience in data analytics and start getting valuable insights within a couple of weeks.