Should I bring my data and analytics to Salesforce? 7 Considerations.
by Jan Hendrik Fleury, on Aug 25, 2021 1:53:20 PM
The share of wallet of the Salesforce platform components in marketing, sales, commerce and service have risen significantly. Salesforce was the first-mover in SAAS CRM based solutions and has been offering value for money from then, adding new functionality and services. However, the time has come for many companies to decide if they should buy even more into Salesforce with components such as Salesforce CDP/Customer 360, Interaction Studio, Datorama, Tableau. As a result, you need to bring your data and analytics to Salesforce too.
People are asking me:
“Should I bring my data and analytics to Salesforce or are there smarter options”.
The short answer is: it depends. In this article, I am sharing seven considerations from my personal experiences. I would like to learn from your experiences and questions. Feel free to contact me to discuss!
1. Data driven maturity
If you are in an early stage of your data driven and CX journey, I can imagine you are satisfied with ‘out of the box’ components that you can compose in a graphical user interface. Probably you don’t exactly know what you want, as of yet. There are many examples of basic implementations with workflows and out of the box intelligence available for inspiration which is handy for a quick basic setup.
At the time where your business grows or when you have gained experiences that are successful, you will want more. Since you are becoming more mature, you want to become more customer-centric, achieve higher ROI on your sales, marketing and service efforts. You are exploring what is possible within the platform. Some functionality can be delivered, some not completely and some not at all. So you have to decide if you can live with missing functionality, or build it, or buy it elsewhere.
The only constant factor is change. You don’t know what requirements you will be having in the near and later future. Consequently, you want to be able to test and learn quickly with point solutions that are not in the Salesforce stack too. Or it is available in the Salesforce stack but not at a reasonable cost. For instance, if you want to integrate with a new point solution, preferably you don’t want to migrate your data (and models) if you swap from ESP, eCommerce platform or a hyper-focused point solution that is missing the functionality that fits with your business. If you centralize your data assets in a centralized unified data platform such as a data lake house you can integrate, test, learn and switch easily.
If you are more of a build versus buy company, the cloud actually enables development and innovation rather than a platform that requires you to only tap into capabilities within their ecosystem. It's an impairment to internal development.
3. Granular control
I have noticed that companies who are becoming more mature, often want to have full control over their data assets including modelling by applying the kind of models they prefer and that this is not limited to the out of the box functionality that Salesforce is offering. Same for identity linking/stitching as another example. The brilliant Salesforce sales teams are selling to the board that ‘everything is possible, buy yourself another module and you are up and running’ and the leadership believes this to be a fact. If this is communicated to data and analytics professionals who actually have to deliver and who want to have control, these personas become nervous. Is it really possible? The answer is that it is not possible to have full control and apply any kind of modelling from the Salesforce platform. Two choices: deal with missing functionality and a big chunk out of the budget (see cost chapter), or build a cloud-based unified data and analytics platform. Recently, I have written an article about shaping a unified data analytics platform.
4. Enhancing the performance of Salesforce components
Having the luxury of a unified data and analytics layer, you are able to enhance the performance of, for instance, Salesforce email because you can apply smarter modelling if you have access to all the data you need and if you can apply the kind of modelling you wish to use. The same goes for modelling to target the right audience for the right add bids based on your first party data and to apply modelling to personalize the website. It goes without saying you can still use out of the box intelligence of Salesforce Einstein when this is offering decent performance. And if the data behind Einstein is richer and of more quality, the performance of Salesforce components will increase. As a result, you will have the best of two ‘worlds’ in one stack.
Note that enhancing Salesforce components may not be best in breed or drive the highest performance. Also, if you want to enhance Einstein's performance then cost is often incremental for another solution.
Is designing and developing an analytics platform expensive? Both developing it yourself (with the help of a service partner) and buying into more Salesforce components is OPEX cost. It is possible to make a business case based on TCO (Total Cost of Ownership). What helps in the advantage of an analytics platform is a factor such as reducing cost of the Salesforce modules. These costs can be massive. Example: if you exceed the number of records that are included in the tier per user, the cost will increase significantly. If you centralize your data outside Salesforce, this will lower the cost a lot in the case where you exceed standard limits. In that case the platform will ingest Salesforce with the data and decisions Salesforce (or another point solution) needs to execute the particular job. Example: why have all your customer records in Service Cloud if it is also possible that Service Cloud only pulls up the records it needs for a case? Meaning actual records. Of course, the result of the job will be delivered back into the unified platform to keep the central view on the data (and analytics as a result of that). An option within the Salesforce world is to store data in Heroku. This has other cons but that is off-topic for this article.
The cost of a cloud platform is mainly based on storage and compute. Storage is cheap. Computing (analyzing / processing for queries and running models) is the main factor but these costs are also relatively low in the cases that I have worked on.
Development cost is another factor to include. This can all be put in a business case.
Also, if you can’t easily move out of Salesforce, there is no incentive for the Salesforce sales team to offer you a good deal. You might find yourself squeezed.
6. Lock-in and Intellectual Property
Aside from cost, most companies want to own the platform data, models, and code. This increases the market capital value of a company. Not surprisingly, CEO’s and owners of a company love that. This is possible in a public cloud platform (aside from the out of the box models in the platform).
You should always make solid agreements about migrating your data (egress) out of Salesforce if you end the partnership. Same for egress of a public cloud platform.
7. Skillset and availability of staff
If you tailor unified data and models to your needs, you need the manpower and the skills to do so. Regardless if you are working in Salesforce or in a public cloud. Knowledge of software development, engineering and analytics is essential to keep up in digitalization. The ease of use of these platforms is rapidly increasing; more and more graphical components are available. Also, as an example, you can apply ML using SQL with Big Query ML so analysts are becoming data scientists. Next to that, a serverless architecture doesn't require the hassle of managing the underlying architecture.
In most cases I see, companies hire service partners and in-house step-by-step in a later stage when they will have become more familiarized with the platform.
There are many factors to consider. Your choices depend on the factors mentioned above and others as well.
What is your take? Thoughts?
Crystalloids helps companies improve their customer experiences and build marketing technology. Founded in 2006 in the Netherlands, Crystalloids builds crystal-clear solutions that turn customer data into information and knowledge into wisdom. As a leading Google Cloud Partner, Crystalloids combines experience in software development, data science, and marketing, making them one of a kind IT company. Using the Agile approach Crystalloids ensures that use cases show immediate value to their clients and frees their time to focus on decision making and less on programming.