3

I am working on a year long Data Analysis project for a moderately sized client, the aim of which is to analyse transactional data and identify for the client its customer profiles, transaction patterns, and assist the business management in making factual business decisions. It has > 1M customers though how many are active, we don't know since we don't have access to their database.

The tools we requested were SQL Server for data storage/extraction, and R for analyses.
However, the client has come back with Access and Excel as the only tools they are willing to provide.

Question: If you were to persuade the IT team that the proposed tools are insufficient for what we have been asked to do, how would you go about doing so?

Following is how I am outlining my arguments:

  • Briefly explain the Data Analysis project deliverables - what amount of data we hope to analyse (we don't know how much space it will take), what sort of data, what will be the outcome, how much will be ad-hoc analyses, etc.
  • Explain how Access will not suit this - primarily due to size limitation for database of 2GB
  • Explain how Excel would slow down significantly if > million rows are added to it for simple manipulation

Although I am planning as per the above, but I feel the IT team may refer me to ways of dealing with Access size limitations, and prevent us from solving this issue reasonably...

Marv Mills
  • 4,424
  • 1
  • 19
  • 27
info_seekeR
  • 133
  • 4
  • Test and research access speeds over the network, as well, especially if you are in a multi-site/WAN environment. – SBWorks May 29 '15 at 05:40
  • 1
    I have edited your question to focus it on asking how to show toolsets do not meet the needs of the project. It is not reasonable to ask for actual analysis of why your particular tools are not relevant for your particular project, but asking how to handle the situation where you don't believe they are good enough is on-topic. If you feel I have not adequately represented your question then please edit it further, but note that questions seeking opinions on tools are off topic and likely to be closed. – Marv Mills May 29 '15 at 10:56
  • 1
    This is an important question that I encounter on a regular basis. I think the key is to represent the requirements that Access/Excel fulfill and compare the priority of those requirements to the performance and resilience requirements. – MCW May 29 '15 at 11:05

4 Answers4

6

There is only one way to "persuade" some party that they tools they are providing are insufficient for your needs. That is- you need to demonstrate the ways that the tools do not allow you to fulfil the requirements of the project. If they counter by showing you ways that allow the tools to meet your needs then clearly the tools are sufficient for your needs, just not your wants.

You have alluded to the fact that the data already exists in a database, you just don't have access. This job appears to be about data analysis so at first sight it is not clear to me (or presumably to them) why you would want to install another database system. If you can extract the data to Excel you can do a large amount of analysis there, but you don't necessarily need to extract the entire database to Excel in order to do this. If you have access to the current database system then just extract what you need for each strand of analysis (or have someone do it for you if you don't have access).

Only once you have exhausted these paths will you know that the tools are not up to the job, and you will have hard proof of your needs. At the moment it is guesswork.

warren
  • 942
  • 2
  • 9
  • 25
Marv Mills
  • 4,424
  • 1
  • 19
  • 27
3

TL;DR

Senior management may say "yes" or "no" to additional project costs, depending on how well you justify those costs and their value to the project. However, the stakeholders and sponsors, not the project team, owns the budget, quality, and schedule risk associated with tool-chain issues once they are made aware of them.

Tool-Chain Justifications and Decision-Making

All tooling and process questions ultimately resolve to an assessment of labor and capital costs as they relate to return-on-investment for a project. Specifically, you should:

  • Identify the cost differentials in terms of time and/or labor between the current and desired tool-chains.
  • Quantify the costs of purchasing, installing, and supporting a new tool-chain.
  • Justify the recommended tool-chain based on concrete deliverables or schedule targets.

Ultimately, any process can be performed with pencil and paper. It just may not be feasible to achieve the desired scope, schedule, or cost targets for the project with a given tool-chain.

It is your responsibility as the project manager to either build tool-chain costs into your project plan, or to justify additional costs as the need for them is uncovered by the project. However, the business decision of whether or not the costs are justified is a strategic decision that rests with the stakeholders (especially the financial sponsors), and making those strategic decisions and any associated trade-offs based on project constraints is their responsibility.

Todd A. Jacobs
  • 50,264
  • 7
  • 59
  • 177
2

Maybe Access and Excel is enough, why not build a minimal viable product with these tools first. Just make sure you structure it (with a good data-access layer for example) so that you can swap to another data-source in a later stage, might Access not scale.

An Agile coach once told me a story about a development team that really thought they needed a Oracle database setup, because of a vision of thousands of users and complex data-structures. The product ended up being used by 3 users, but still they had to pay large sums of licensing fees to Oracle to use the product. While Excel/Access with some macros would have been enough...

Maybe costs is the reason for pushing for Access. SQL-server needs a server and SQL licenses, which are not so cheap. Maybe you can suggest to use alternative SQL servers like MySQL on Linux to prevent costs.

warren
  • 942
  • 2
  • 9
  • 25
Niels van Reijmersdal
  • 5,553
  • 20
  • 32
  • 1
    SQL Server Express has a limit of 10GB for a database. That's 5x better than Access, at least – warren Jun 04 '15 at 21:21
2

In a perfect world, the answer is:

  • Perform a requirements analysis
  • Write down your assumptions
  • Discuss the pro and cons with your Stakeholder

As others advised before, try to go this way. But be aware that it ignores the human afraid of change.

Expand the list above by a risk and chances analysis, e.g. SWOT. This should help to bring a new view to the discussion by addressing the need for change topic.

Regardless of the result, document the assumptions that brought the decision.

Tob
  • 1,767
  • 1
  • 12
  • 24