Introducing Service Level Expectations to Software Development

Introducing Service Level Expectations to Software Development

A question I receive frequently is, “What’s the difference between building software and building a service?” When answering this question, I often think about the simple difference of cooking in your kitchen for yourself in contrast to cooking in a restaurant for others. In this instance, building software for internal use is like cooking in your kitchen, whereas building software for externalization is like cooking in a restaurant. Looking at the two, both have the same inputs--ingredients--and output--food, but the primary difference is building a service requires service level expectations while building a software does not. Restaurants have a lot of standards, regulations and processes in place that even the dream home kitchen doesn’t need to grapple with.

?Service level expectations, or what your customers expect from your service, include utilization, availability, latency, user experience, and security, among others. For a restaurant, this translates to the ease of booking a table, the responsiveness of staff, the wait time for food and drinks, and so on. Meeting or exceeding service level expectations requires a complete system of highly synchronized mechanisms operating behind the scenes with strict processes and the perfect blend of creativity with repeatability and precision.

?While software doesn’t inherently have service level expectations, we still apply this service mentality when building both internal and external software, evidenced by our working backwards approach. By establishing and clearly communicating these service level expectations when building software, we not only maximize our chances of exceeding client expectations, but also improve the experience for our developers internally.

From start to finish, this translates into having clear criteria for when the service is generally available, setting clear service level objectives and communicating how you do or don’t meet them, having well-designed APIs, never breaking compatibility, protecting against erratic traffic, and so on. This mindset also informs our platform-based approach to development at Goldman Sachs, which focuses on designing services that are compatible and highly reusable across teams, so we can increase velocity and operate more efficiently.

Leveraging this approach, we have built several platforms, for internal and external use, in the last few years. This platform-first approach has enabled us to develop cloud-native platforms that serve as a foundation for teams to build on top of, so we can increase velocity and get to market faster to meet client demands. This foundational, flexible technology stack has allowed us to go from building platforms to increase efficiencies internally, to externalizing those for client use, to now being a service provider of embedded finance.

The success of this approach can be seen across our work at Goldman Sachs with platforms like Marquee, Transaction Banking (TxB) and our cards business, which we have been able to continually expand and scale in an innovative, efficient way. For example, by building TxB from a clean sheet of paper with no legacy technology debt, the team was able to get to market fast and bring clients a fully cloud-based digital platform that’s user-friendly, operates in real-time, is API-centric, and uses the latest data technology to provide analytics to clients.

From the launch and expansion of Visual Structuring on Marquee, the digital platform for institutional and corporate clients of Global Banking and Markets, to our developer-centric businesses in Platform Solutions, we’ve been able to bring service level expectations to software products and efficiently address the specific needs of our wide range of clients through digital-first solutions that abstract complexity and help them operate more efficiently, grow their businesses, and best serve their end customers and clients.

?Applying the same service mentality to building software as one applies to building services increases the likelihood of exceeding client expectations. Establishing, communicating, and meeting clear service level expectations provides us with a framework to deliver software, both internally and externally, that is best positioned to facilitate positive outcomes for all involved.

Terence Bennett

Chief Executive Officer @ DreamFactory Software | API Generation, Data Access Automation

1 个月

It's really interesting how you apply a service mindset to software development at Goldman Sachs. How has this approach impacted the developer experience and client satisfaction in your projects? Would love to hear more about any real-world examples that highlight these benefits!

回复
Linda Goodman

Community Chief of Staff | Program Director at GLEAC

4 个月

It’s not just about building—it's about building with intention, focused on real-world outcomes.

回复
Daniel Terry

Senior VP Engineering - Asset & Wealth Management Division

1 年

Love this

回复
Alessandro Petroni

CEO - Executive advisor | GTM strategist | FinServ specialist | Open Source leadership | publisher | ex-Red Hat | ex-Merrill Lynch | ex-Deutsche Bank

1 年

Interesting reflection, Marco. Software is a just a tool that is part of a solution which is expected to provide value. However value is seldom mapped to just software features and capabilites, which are enginnering outcomes. Thus the of need product and service managers, educators, documentation writers, support people and consultants to unlish and to continue increase value of solutions, their use, and the supporting software.

Maciek Samsel

Architect at Amalgamated Bank (contract)

1 年

That is the fundamental difference in creating software in industrial operations technology and technology entrepreneurship companies. Goldman Sachs vs. Amazon. And that has been there at Goldman for long time in many teams, but some missed that.

回复

要查看或添加评论,请登录

Marco Argenti的更多文章

  • Leadership

    Leadership

    Last week, we introduced our newest partner class at Goldman Sachs. As we minted our next generation of leaders, I…

    9 条评论
  • Rules and Thermostats: Balancing Risk and Reward in the AI Era

    Rules and Thermostats: Balancing Risk and Reward in the AI Era

    There is a balance in the risk and reward between following the instruction manual and creating your own new design or…

    3 条评论
  • Nature and Nurture: The Coinciding Evolution of Humans and AI

    Nature and Nurture: The Coinciding Evolution of Humans and AI

    When people ask me what differentiates generative AI from other emerging technologies and how it will continue to…

    69 条评论

社区洞察

其他会员也浏览了