It’s been a while since Robert and I got together on the ole’ podcast, so we were both eager to dive into this episode. And today we tackle one of those activities that is part of every implementation. The mid-morning cookie break? No! Data import, of course!

Robert and I discuss the strategies to employ and the pitfalls to avoid when planning your data migration.

Getting the right data cleanly into the system for launch is a requirement for successful adoption. And giving users the right data when and where they need it to do their job is critical. After all, that’s kind of the point of the system, right?

But this discussion is really about avoiding the pitfalls we’ve seen too many times. Trying to load tons of history, allowing a bunch of duplicates in, or just generally letting junk into the system. These are bad ideas. While data migration it is such an important part of every project, it can be a daunting task. So today we’re going to outline some things you can do to ensure a smooth data load.

1. Know Your Data
Buddy up to your data early in project planning. You’ll want to really have a handle on all the sources, how they connect and overlap, as well as what kind of shape they’re in. All too often this doesn’t happen up front and the amount of effort involved in data migration is underestimated and throws schedules off track late in the project cycle.

Two areas you should focus on:

Data Volumes: Know the volume of data you are dealing with. We’ve pointed out in our performance podcasts the impact volume can have, so if high volumes are expected, then you may need to plan more time for performance testing and tuning.

Data Preparation: Strive early on to get an idea of the cleanliness of your data. Will it need to modified to conform to crm od standards? Do you plan to de-duplicate or enrich your data?

For each data source, and record type, estimate the LOE for prep activities and build it into your plan. It takes time – data imports can take minutes, hours, days or weeks. It depends on volume, number of fields and the type of records being imported. Use your test data as benchmarks to establish some expectations, and reconcile with the project plan.

2. Less is More
When you are crafting a data migration strategy, you have to think about it from the user’s perspective. Picture them. In fuzzy bunny suits…What do they need to do their job? How can we make it easy for them to find what they need without adding a lot of noise?

Anyway, the principle of least privilege dictates that users get access to the data they need to do their daily jobs and no more. In some cases that might mean they need access to only a few hundred customer accounts. Some cases may call for access to a whole lot more.

The important point is to ask the questions and justify the data for real business reasons. If you are just dumping everything because that’s most convenient, you’ll likely be dealing with the complaints later.

3. Align data with your business objectives
Since data is such an important adoption driver, it deserves a good amount of attention in your deployment. All too often data import is handled as a separate project, even though it is usually a critical path for go-live.

Having import silo-ed from the rest of your project can often end up with all the data getting dumped from one system into another, with negative consequences for users. What to do? Well, make sure the data import supports your business objectives.

Easy to say, but what does that really mean? Here’s an example. A company objective for crmod is to have provide simple, concise screens which will drive user adoption. BUT the account data we have is 50+ fields. Way too much.

We review the data and determine that we can eliminate 10 fields and push additional fields into related items (Assets, activities), thereby simplifying the Account screen and improving chances for adoption.

Same goes for volume. Do the business objectives call for accessing 5 year-old activity records? No? Then why import ‘em?

4. Quality matters!
Also known as “garbage in / garbage out”. Don’t underestimate the importance of high quality data. Users faced with a bunch of junk or redundant data are far more likely to be frustrated with the system, be reluctant to contribute information to it, or abandon it altogether.

The initial data load is a great opportunity to address any inherent data quality problems. All too often we’ve seen customers decide that they “don’t want to risk throwing anything out” – sounds like my wife!

But that’s a personal problem. For CRMOD customers, I would always weigh adoption and ease of use more heavily than the risk that 3 or 5 years of activity data might contain a gem. If it’s a real concern, dump it into a local database and make it available offline.

And there are great services out there – like our partner ActivePrime – that create tools for helping identify duplicates and enrich your data. These can cost a few bucks, but might be well worth it.

5. Keep your room clean
When I was a kid, eventually my room would get so messy my mom would kick me out of the house for an afternoon and I’d come back to find it neat and tidy – and probably missing some old toys. After that, it was a constant refrain of “keep your room clean”!

What does this have to do with data migration? Not sure.. Oh, wait – I’m saying that now that we’ve devoted all this time and energy to giving the users a nice tidy data set, we want to keep it that way! Wishful thinking!

Yeah, that’s what I told my mom. But you’ve got to try! Seriously, a good data strategy isn’t all about initial migration – that’s only a small part. A good data strategy keeps an eye on the system, employs business practices – like field validation rules – aimed at keeping data in line. And involves users in data maintenance, like allowing users to flag records for merge or deletion.

Episode Wrap-up
Boy, we could talk for hours on this topic. Data loading was such a big part of every project when I managed consulting – we had dedicated full-time resources and an offshore center to manage it.

And we haven’t really even touched on the technical ins and outs of a successful data migration. I mean, there are so many things to talk about, including data cleansing approaches, data loading mechanisms, and on and on.
Yikes. Sounds like a few more podcasts on the horizon…

That’s all for now, see you next time!

Share and Enjoy:
  • Digg
  • LinkedIn
  • Facebook
  • Mixx
  • Google Bookmarks
  • Twitter
  • StumbleUpon
  • Technorati
  1. Damian Hoy on 17 Sep 10 1:06 am

    Great podcasts guys! Very entertaining and informative.

    I will keep these strategies in mind during my next implementation.

    Keep the podcasts coming.