OnCrawl Labs is our playground. As part of our R&D we ask: “What SEO tasks are too expensive, or too incertain, or too time-consuming, or too complex to be carried out correctly today?”
We use today’s tools — like machine learning, deep learning, and data science — to shape a solution.
And as soon as we have something we know can help you do better SEO, we share it. Even if it’s not something we can include in a one-size-fits-all product, we still want to give you ideas, to help you get started, and to make you think outside the box.
In this article we’re going to talk about what you need to get started with OnCrawl Labs.
Input data for OnCrawl Labs projects
Data SEO depends on the input data, that is, the data you have to start with. For example, if you’re trying to make predictions, it’s important to look at significant events and at what things normally look like in order to project that into the future.
You might have work to do to obtain or to format your data for the projects in OnCrawl Labs. Depending on what you have, and how and where it’s stored, this process can look very different from our process, or from your competitor’s process. It’s extremely personal.
In each OnCrawl Labs project, we let you know what you need to prepare or provide before starting, and in most cases, we also provide examples or sample input datasets.
Running and adapting the project
Next, you’ll need the appropriate tools and environment–and people–to run the project scripts. Tools might include programs. For projects using deep learning, tools might even include physical computers with enough memory to run computationally-heavy processes. We’ll let you know what you need right from the start.
In some cases, you’ll need to adjust variables, account for specificities in your data or your industry, or make other changes to our base project.
This is, again, unique to each company and to each project.
In OnCrawl Labs, we’ve provided you with the steps to make the project work and the base code that we have rigorously tested and documented. This code can easily be copied, adapted or transformed by your own teams, based on your own needs.
The output data, or the results of the process that we offer, can be adapted to be a perfect fit for whatever you need to do with them next.
For example, you might need to change their format in order to easily feed them to the next step in your process. Or maybe your data analyst wants to create visualizations. They might use software that requires a specific import format.
The list is endless, depending on what you intend to do with the results, who needs to see them, and how they need to be able to interact with them.
Finally, the work of analyzing this data and making decisions based on it is probably not something you want to leave up to the machines… for now.
It obviously depends on that original SEO task that was too costly, too complicated, or too uncertain to handle normally. If your problem dealt with uncertainty, you might need to be able to summarize the data and the results for decision-makers; if it was a problem of scale, you might need to finish implementing the changes on your website. Or you might just need to provide the results as-is to the person who needs to copy and paste them, and then track the impact of your changes in Google Analytics, for example.
In OnCrawl Labs, we give few suggestions as to how the results can be used to change your SEO process, or how they can inform your decisions. In general, we try to help you think about the next steps to get the most out of each project.