In short
- Migrate the method, not just the numbers. The method is what makes next year comparable.
- Rebuild the most recent closed year first and reconcile it category by category, never on the total.
- Ask for the handover pack while the relationship is still warm. It is far harder to get later.
The risk in this move is not the software. It is comparability. A consultant-produced footprint carries decisions inside it that never appear in the report: a boundary, a base year, a factor set, a set of estimates. Rebuild the numbers without those decisions and your first in-house year will differ from the last consultant year for reasons nobody can explain, which reads as a restatement rather than a reduction. So migrate the method first, rebuild the most recent closed year in the new system, and reconcile it to the published figure before you trust anything the platform tells you about this year.
What do you need from your consultant before you start?
Ask while the relationship is still good. A handover request sent after notice gets a very different response than one sent alongside a renewal conversation, and most of this is yours anyway.
| What to ask for | Why it matters next year | If it does not arrive |
|---|---|---|
| Activity data by category, with units and source | It is the input to every future cycle | Rebuild from your own source documents, most recent year only |
| The factor set: database name and version | Without it, your trend mixes methods silently | Adopt one named set going forward and document the switch |
| The boundary and consolidation approach | Decides what is even in scope | Write it from the report and ask them to confirm by email |
| Base year, and any restatements made since | Every target and trend hangs off it | Treat the earliest fully documented year as your working base year |
| Exclusions and the materiality justification | Auditors and tender evaluators ask | Redo the screen before your next submission |
| Assumptions and estimation methods | Where the judgement actually lives | Re-derive them and have them reviewed once |
The third and fourth rows are the ones people forget to ask for and the ones that cost most to reconstruct. Ask for them in writing even if you never open the spreadsheets.
How do you stop the baseline breaking?
Run the old and the new against each other before you rely on either.
Rebuild the most recent closed year, not the base year. It has the freshest evidence and a published figure people still remember.
Compare category by category. Two totals can match while four categories are wrong in opposite directions. A total-level match proves nothing and has misled many first migrations.
Set a tolerance before you look at the numbers. Deciding what counts as a match after seeing the result is how a variance becomes a rounding difference.
Write down every difference above the tolerance and its cause. That document is the bridge between the consultant era and the platform era, and it is the first thing an assurance provider will ask for.
What do you do when the numbers do not match?
They will not match exactly, and that is normal. What matters is whether you can name the reason. Five causes cover almost everything.
A different factor version. The same activity, a newer database. Explainable, and the cheapest kind of difference.
Currency or unit conversion. Spend converted at a different rate, or distances calculated by a different method. This is the difference most often mistaken for an error.
A boundary difference. A site or entity in one version and not the other. Fix the boundary rather than the number.
An allocation or estimation choice. The consultant apportioned something and you apportioned it differently. Pick one, document it, and apply it to both years.
An actual error, in either direction. Rebuilds find mistakes in original work reasonably often, and they also find mistakes in the rebuild.
Then choose one of three outcomes for each difference: adopt the new number and restate with a published note, keep the published number and carry the difference as a documented bridge, or correct the rebuild. What you must not do is adopt a new number quietly. A silent restatement is the single fastest way to lose credibility with a customer who kept last year's answer.
Which year should you move first, and what about the older ones?
Start with the most recent closed year, because it is your reconciliation anchor. Then decide deliberately whether to backfill.
Backfilling is worth it when you publish a trend, hold a reduction target, or bid into a scheme that scores year-on-year improvement. It is not worth it when your obligation is a single current-year figure for a customer questionnaire.
If you do backfill, do it with the factor set and boundary you have adopted going forward, and label those years as recalculated. A consistent recalculated series is more useful than a mixed one that happens to match what was published.
What should stay with the consultant?
Less than before, and not nothing. The judgement work does not become easy just because the routine moved in house.
Keep them for boundary changes after an acquisition or a divestment, for modelling a newly material category, and for the first submission into anything regulated or assured. Complex purchased goods work is a good example, which is why scope 3 consulting exists as a service rather than a feature. Product footprints stay a project regardless of your platform: LCA, EPD, MKI and PCF work is consultancy with its own standard and its own reviewer.
A useful arrangement is a short method review at the start of each cycle rather than a full engagement: the consultant checks the boundary, the factor set and any new category, and you run everything else. That is what carbon footprint consulting looks like once you own the inventory.
What does the first in-house cycle actually feel like?
Heavier than the second one, because work that used to happen inside somebody else's hours now happens in your calendar.
The chasing is the part that moves. Your consultant was quietly absorbing the effort of getting twelve months of records out of six people, and that effort does not disappear when the invoice does. Name one owner with protected time and a data owner per category before the cycle starts, not during it.
A small business reviewer said on G2 in August 2026 that it takes a lot of manual labour to load data, and that once the data is there it works perfectly, but getting it loaded is the challenging part. That is the trade you are making: an invoice becomes a calendar commitment. It is usually a good trade, and it should be made with open eyes.
What does Hedgehog do for a migration, and what are the limits?
The platform is built for the case where somebody else did the first footprint. You can add your own organisation-specific or supplier-specific CO2 data alongside more than 20,000 spend-based and activity-based factors, which is how a consultant's supplier figures survive the move. You can collaborate with colleagues, consultants and auditors in the same account, inviting them as data owners, auditors or managers, so the outgoing consultant can check your rebuild without holding the keys. An AI guide walks through GHG Protocol setup and helps identify your data sources, data owners and documents, and human GHG experts are reachable in-app. A free account needs no sales call, and Pro starts at EUR 1,200 a year.
Two limits that bear directly on a reconciliation.
Traceability is not complete. A mid-market reviewer said on G2 in July 2026 that they could not see the conversion factor applied when their input in one currency became an entry in another, or which calculator produced the distances in their distribution overview. During a migration that is exactly the visibility you want, so test it against your own hardest category before you commit.
Few integrations today. A mid-market reviewer on G2 asked for more integrations in future, in June 2026. If your reconciliation plan assumes an automated pull from a finance system, check it first.
What should you do first?
Send one email today asking for the six items in the table above, before any notice period starts. Then rebuild a single category of your most recent closed year and reconcile it. One category tells you how big the whole exercise is, and whether the handover pack you received is actually complete.
You can do that on a free account without a sales call. If you would rather have someone look at your existing deliverable and say which parts are safe to carry across, book a meeting.
Sources: Hedgehog platform, Hedgehog on G2. Verified 27 August 2026.
Facts on this page were last verified on 2026-09-17.
