At Programmai we have a simple mission to turn your complex customer data into online marketing success. Our predictive platform tells you which consumers to target based on propensity to purchase and future lifetime value in order to deliver optimal levels of investment and efficiency.
One common concern we get is why a company should outsource their online predictive marketing versus building it in house. The answer varies a bit depending upon whether you already have a Data Science team up and delivering business value or not, however there is quite a bit of overlap.
Already have a Data Science team?
The most important point to emphasize is that Programmai is not a Data Science team replacement. We don’t perform the full gamut of services from customer and category analytics, to helping with supply chain issues, or aiding in pricing decisions. We focus on helping customers improve their online marketing efficiency so they can increase revenue, decrease costs or both.
There are so many ways Data Science can provide business value that companies need to consider the strategic benefit of each project they take on. The build versus buy decision should not be applied to Data Science as a whole, but to each individual way it can add value to the organization.
With each activity ask yourself, “Will this give me an advantage over my competitors?” and “How many of my competitors are already doing this or are planning to do this?”
Data Science teams should focus on those activities that are strategic differentiators. Namely, the ones that add a lot of customer value while having few competitors performing them. Table stakes are those activities that customers expect you to be performing and all your competitors are already doing. Consider whether it makes sense for your in-house team to be focusing on them.
Strategic differentiators are those things that your competitors aren’t doing but should be because of the value your customers would receive from them. For instance, Zalando runs a highly optimized predictive supply chain to minimize delivery times and delight and surprise customers with how quickly their packages arrive. I worked on a project to cluster makeup colour preferences onto a geographic map so that acquisition campaigns would target potential customers by their local shade preferences. These are the types of projects you should be investing resources in as they provide the best opportunity to grow the business.
Table stakes are those activities that all your competitors are doing and customers just expect you to do and do well. Your website visitors know that platforms such as Facebook and Google offering very granular targeting abilities and that you can see where and for how long they were on your site. So when a visitor spends 10 seconds looking at a single product and leaves they can become frustrated and annoyed when that item follows them around the internet.
Companies tend to invest a lot in these more common areas, which is ironic since they provide the least strategic value as everyone else is doing them too. They also offer the biggest economies of scale and therefore saving opportunities if they were outsourced.
Even for companies with Data Science budgets well into the 7 or 8 digits, Programmai can offer great value by taking over those activities that distract from delivering true strategic value, saving time and money.
Thinking about or are starting to build a Data Science Team?
Starting a new team has all the issues, involving focus versus value, described above for existing Data Science teams and more. The difficulties come from the need to create capabilities in several areas, while also building a data-driven culture change throughout the organization.
Onboarding can be especially challenging in Data Science due to the number of things new starters need to learn to be successful. This means that most new team members go through a phase of negative productivity while they learn the company’s business, its data and the tools and techniques that have been (or worse need to be) agreed upon.
New starters on a Data Science team go through an especially deep and long period of negative productivity increasing the risk of teams running out of business goodwill before they start showing value.
Often during this phase, the business will start to question the decision to invest in Data Science when they see money going out the door for expensive talent, software and hardware to perform activities such as building modelling pipelines, which create value but has little visibility to the customer. The worst part of this phase is that the amount of time it takes to become a productive team isn’t known and can drag on far longer than anyone initially expected or budgeted for.
Another common issue when building a Data Science team is to focus solely on delivering the output of a process like a predictive model, such as purchase propensity scores. Often they forget that to be actioned they need to be integrated into a platform, such as Facebook, which requires knowledge of APIs, complex permission configurations and of Facebook’s own product offerings. Additionally, they will need knowledge of campaign and marketing strategy to be able to successfully turn those scores into business value. This can get very complicated and requires experimentation, further delaying the team’s ability to show value to the business.
Successful Data Science teams need to show business value which means more than just delivering predictive scores, but also correctly integrating and actioning them.
New Data Science teams need to show value quickly. Outsourcing table stakes work to a company like Programmai allows them to do this, at a fixed cost risk, while giving them the time required to properly grow and nurture an in-house team.
Some teams have expressed a concern that outsourcing any work will open the door to management contracting out all the work. To those, I would say that the message to the executive team needs to be that Data Science is a source of strategic competitive advantage but not all Data Science work is equally important. Build a roadmap of the activities you could be working on and place them on a matrix of customer benefit versus competitor progress and focus strategically.
Our goal is to be an outside asset that uses the sophisticated predictive platform we’ve built to pick up those projects that are of the least strategic benefit to you, but which your customers nonetheless expect you to excel at. Freed from the burden of these activities, your team can focus on driving real change through the organization and making Data Science a business priority. We know from experience that this isn’t easy, so if you need outside support we would be happy to meet with your senior stakeholders to explain the benefits of outsourcing the table stakes workload to companies like Programmai.