The right Teradata migration partner is the one that can prove, on your own workload, that it understands what your Teradata system does today and can deliver the same business results on the new platform. Brand, slide decks and conversion-tool demos tell you little about that. Ask for evidence, test the team on your own code and award the large contract only after a paid pilot.
We at DWHPro are one of the firms you might evaluate, so read this as a buyer’s checklist written from the delivery side. It is the checklist we would want our own work judged by.
The three questions that matter most
- Do they understand the source system? Most migration risk sits in what Teradata does implicitly: performance built into the physical design, workload management rules, data integrity and business logic inside the database. A partner that knows the target platform well but treats Teradata as “just another database” will discover these things late and expensively.
- How will they prove the results are the same? Ask for their validation method in detail: what is compared, at which levels, who signs off and how differences are investigated. A partner that talks mainly about conversion speed has not answered the question.
- Who will actually do the work? Ask for the named senior people, their share of the effort and how long they stay on the project. The experience shown in the proposal is only useful if those people are on your project.
Questions that separate experience from slides
- How do you migrate while Teradata stays in production and data keeps changing?
- How do you validate billions of rows without relying on row counts?
- What do you do when a critical batch runs slower on the target than on Teradata?
- Which parts of our code would you expect to convert automatically, which manually, and which could silently produce different results?
- Which workloads would you run unchanged through a virtualization layer, which would you convert, and which would you redesign, and why?
- How long do you recommend running both platforms in parallel, and what are the exit criteria?
- What happens if the cutover fails?
- When would you advise us not to migrate?
Experienced partners answer with concrete examples and trade-offs. Inexperienced ones answer with methodology diagrams.
Get the complete Teradata tuning book, free
Teradata Query Performance Tuning: The DWHPro Guide
By Roland Wenzlofsky & Artemios Vogiatzis
206 pages on statistics and demographics, access paths and indexes, query execution plans, and joins. Get the full PDF and practical Teradata updates from DWHPro.
Confirm your email to receive your download link. Privacy policy. Unsubscribe in any newsletter.
About the free bookTest them on your own workload
The most revealing step costs little: give every shortlisted partner the same sample of your Teradata code and processes, including a difficult stored procedure and a real batch chain, and ask them to explain what they see, what they would change and what could go wrong. You will learn more in an afternoon than from weeks of proposals.
Before awarding a large migration contract, commission a paid pilot on a representative, demanding part of the warehouse. Judge it on validated results, measured runtimes, consumption on the target and the quality of the documentation, not on how many objects were converted.
Warning signs
- The proposal is built around an automated conversion tool and says little about validation.
- A fixed timeline and price are offered before anyone has analysed your workload.
- No detailed reconciliation method, or a reliance on row counts.
- No answer to “when should we not migrate?”, because every assessment ends in a migration recommendation.
- Senior experts appear in the pitch but not in the staffing plan.
- No plan for parallel running, rollback or the knowledge transfer to your team.
- Savings are promised as a percentage without stating the assumptions behind them.
Large integrator or specialist?
Both can be the right choice. Large integrators bring delivery capacity for very large programmes and many parallel workstreams. Specialist firms bring senior people who work hands-on and know the source platform in depth. Many organisations combine the two: a specialist for assessment, design, the difficult workloads and validation, with broader capacity for volume work. Whatever you choose, apply the same questions to every candidate.
Commercial terms worth insisting on
- A fixed-price assessment and pilot before any large commitment.
- Named senior staff with agreed availability.
- Acceptance criteria based on validated business results, performance and cost, agreed before delivery starts.
- Transparent handling of scope changes.
- Knowledge transfer throughout the project, so your team can run the platform afterwards.
- A defined stabilisation period after go-live.
How DWHPro works
DWHPro is led by Roland Wenzlofsky and works with a network of senior specialists. We have worked with Teradata for more than twenty years, from performance tuning to physical design, and support migrations to the cloud. Our approach follows the checklist above: we start with a fixed-price assessment at a published price, prove the target design in a pilot on a demanding workload, validate business results rather than object counts, and say so when optimising Teradata or moving to Teradata Cloud is the better option. Our published Swiss retail case study describes a Teradata-to-Snowflake migration delivered this way; named references are available under NDA for qualified evaluations.
See our Teradata-to-Snowflake migration services for scope, deliverables and prices, or start with an independent platform assessment if the target is still open.
Common questions
Should the migration partner also be a reseller of the target platform?
It can be, but ask how that affects their recommendation. An independent assessment is most useful when the option of staying on Teradata, or choosing a different target, is genuinely open.
How many partners should we evaluate?
Enough to compare approaches, few enough to evaluate properly. A short list of three, tested on the same sample of your workload, usually gives a clear picture.
Is a paid pilot worth it if we have already chosen a partner?
Yes. It is the cheapest way to replace assumptions with measured effort, runtimes and costs before the large budget is committed.
Planning a Teradata-to-Snowflake migration?
Review the assessment, conversion, reconciliation and cutover work needed for a controlled transition.
Get the complete Teradata tuning book, free
Teradata Query Performance Tuning: The DWHPro Guide
By Roland Wenzlofsky & Artemios Vogiatzis
206 pages on statistics and demographics, access paths and indexes, query execution plans, and joins. Get the full PDF and practical Teradata updates from DWHPro.
Confirm your email to receive your download link. Privacy policy. Unsubscribe in any newsletter.
About the free book