Data Migration to AWS for Consumer Products & Services Firm
- Custom Software Development
- Confidential
- Quality
- 3.5
- Schedule
- 4.5
- Cost
- 4.5
- Willing to Refer
- 4.0
"ScienceSoft has very solid project management."
- Consumer Products
- Portland, Oregon
- 201-500 Employees
- Phone Interview
- Verified
ScienceSoft facilitated a migration from a legacy system to Amazon Web Services. They also helped implement best standards for hosting, databases, and servers.
Following the migration, the amount of throughput doubled. ScienceSoft delivered efficient, transparent project management and scrum-style updates. The team communicated clearly, reguarly, and concisely, and they were quick to respond to feedback and fix issues.
A Clutch analyst personally interviewed this client over the phone. Below is an edited transcript.
BACKGROUND
Please describe your organization.
At the point in time that we worked with ScienceSoft, it was called something else, and it offers a cashback portal to customers. A customer may affiliate himself or herself with the portal, make an account, and then when they proceed to follow links and shop at their favorite brands, like Wal-Mart or Amazon, they can earn cash back through what's normally termed the affiliate marketing. We then share those profits with the customer. That is our core portal and platform.
What is your position?
At the time of the ScienceSoft integration, I was vice president of engineering, but now I'm the CTO.
OPPORTUNITY / CHALLENGE
What business challenge were you trying to address with ScienceSoft?
I inherited the contract. They were about a month into it. The original idea was to facilitate a migration from the prior hosting to Amazon Web Services hosting and, as such, to implement some of the best standards for hosting, databases, and servers.
We haven't since expanded the scope of their involvement, mainly because the system that they were migrating, and assisting with migration, has been slated for sun setting, and a new system will be replacing it. Their expertise was no longer required with the team that we had assembled for the new system. Essentially, with the additional staffing in-house, we now have the talent internally.
SOLUTION
Please describe the scope of their involvement.
We were allocated a team of four developers. Two were working, one was the account manager, and one was the project manager. Our interactions were mediated by the project manager. Our developer was working with their developers when they were on specific things, like moving a specific service, to understand the implication.
How did you come to work with ScienceSoft?
I'm actually unaware of that. That was done prior to my start, in a period of time when there was no CTO or vice president of engineering. It was kind of an interesting period for the company. I'm not actually sure about the selection criteria. If I had to put a hunch on it, it was most likely a personal reference from somebody. Most of the decisions seem to have taken that route.
Could you provide a sense of the size of this initiative in financial terms?
I'm not sure. I know it was hourly, and there were a few extensions. Obviously, the costs were beneficial, compared to a United States-based consulting agency, due to the currency situation, but I do not recall the actual financials. It was an hourly-based project, so their statements of work were basically estimatations for the amount of time that it would take. It was a fixed price but based upon hourly.
What is the status of this engagement?
The relationship has ended, due to the implementation of a new system and the addition of in-house staff.
RESULTS & FEEDBACK
Could you share any statistics or metrics from this engagement?
They were able to increase throughput approximately twofold. That's not saying too much, however. There were signs that this was one of their first gigs on an Amazon migration, merely from the way that they configured the environment. They didn't follow some of the best practices for security and for performance speed. The way that they configured the databases in the replica, the server access, the auto-scaling isn't there. However, a lot of that was hampered by the old codebase that they were having to migrate. It did not lend itself to Amazon's scale. It was not designed for a cloud's scale, and they did not have the scope or the time to re-architect the system.
It was to basically take a giant, monolithic, old stack and place it into Amazon. They were successful in that and they were able to make out approximately a two times throughput compared to the old system, by load balancing the few components that would allow them do to that. It's kind of a hard one to judge them. The stack does not lend itself to cloud scale. However, there were some of the small things that made it clear that they were not as experienced with Amazon as they perhaps sold themselves to be. Then again, I was not involved in the selection process, so I don't know how they presented themselves and their services. If they sold themselves as experts, or if they sold themselves as experts on PHP and MySQL and they were willing to incur the learning cost of Amazon, I'm not sure.
Since we have terminated the contract, we have fixed malfunctions in-house. We did bring them back on for a very short spike when they needed to modify the database backup solution that they had elected. That's another weird one, Amazon does that for you, and you don't need to pay somebody to do that. There are some examples of things like that, but they were able to come back in, they were able to bring on the same team members who had been involved beforehand. They were able to resolve that little bug quickly and get it up to date, and it's been running smoothly since then.
How did ScienceSoft perform from a project management standpoint?
They delivered the majority of their milestones on time and within budget. They're good at communication as well. We used JIRA for managing the project. It was our in-house JIRA, and we created the project specifically for them. We appreciated how flexible they were with that.
What distinguishes ScienceSoft from other providers?
Nothing that stands out. ScienceSoft has very solid project management, and they do very well on the communications, which shouldn't be a cause to stand out, but it actually is. I think they can be praised for that quite a lot. They were always very clear and concise on where they were. There was never that offshore consultancy dead air or silence that is often so dreaded. They're very on top of it with their communications and their scheduled communication, both written and in a verbal scrum-style updates. They did very well in that respect.
Is there anything ScienceSoft could have improved or done differently?
No. They did quite well both in written and verbal communication. They were on top of it. Their estimations were close. In software, there are always some surprises, and when those occurred, they were quite quick to prompt the conversation for a re-estimation based on newly learned insight.
Without being involved in the selection criteria, I don't know how well they pitched themselves as Amazon experts or if they did at all. If they had pitched themselves as Amazon experts, I would say that they need to reevaluate what that truly means. However, if they were forthright and it was a good learning opportunity, and they were more about being PHP MySQL experts, then they did a great job. They actually got it into place and it's functional. It's just that it doesn't follow some of the best practices with Amazon.
When working with ScienceSoft, I would say it's important to have adequate internal technical oversight, and use ScienceSoft in more of a staff augmentation capacity, rather than in a project bid consultancy style. For a migration like this, it would have been beneficial to have an internal insight that had the knowledge, but didn't necessarily have the time to do the work. Whereas, perhaps in other components of ScienceSoft's expertise, they could absolutely take a project 100 percent and run with it. In this case with Amazon, I would say that they could have done more staff augmentation. If they had guidance, it would have been easier and better for them.
RATINGS
-
Quality
3.5Service & Deliverables
"They did a lot better than most people. They're up in that 75 percent, but I know there are rating mechanisms where giving anyone less than a four is like a kiss of death. They did a good job– above average for an offshore partner."
-
Schedule
4.5On time / deadlines
"Any of the slippage, honestly, I wouldn't even say it was their fault because it was things that were discovered as they began the migration...We moved the schedule by about two weeks but, honestly, that's even a miracle with what they had to deal with."
-
Cost
4.5Value / within estimates
"There wasn't any scope creep that wasn't clearly founded. It didn't seem like there was any of the bloat or the overflow that you sometimes get. They did a very good job of getting it done."
-
Willing to Refer
4.0NPS