An operator spends $150K–$400K on a workover — a rod job, a re-perf, a chemical treatment, a pump swap — and production ticks up the next month. Everyone nods. But the honest question is harder: did the well actually gain barrels it wouldn't have produced otherwise, and is that gain going to last?
The trap is judging a workover against last month. A well on a 40% annual decline was going to fall regardless. If you measure the uplift from the pre-job rate without accounting for the decline you avoided, you'll flatter every intervention. And if you only look at the first 30 days, you'll miss the ones that spike and roll right back over. Here's how to measure it against the record instead.
Start by rebuilding the pre-job baseline
Before you can score the uplift, you need to know what the well would have done without the intervention. Pull the well's monthly production history for the 12–18 months leading up to the workover date and fit the decline. That gives you the trend line — the rate and the slope — that was already baked in.
Ask the Wellsite data lake for the well's decline rate and monthly oil and gas history through the month before the job. Now you have a forecast: if nothing had changed, the well produces roughly X bbl/month and falling at Y% per year. That forecast is the bar the workover has to clear. Anything above the projected decline curve is real uplift; anything at or below it is just the well doing what it was already doing.
Score the uplift, not the bump
With the baseline in hand, the arithmetic gets clean. Take the actual post-job production and subtract the projected baseline, month by month. The cumulative difference is the incremental oil (or gas) the workover bought you.
A concrete example. Say a well was making 900 bbl/month and declining toward a projected 820 the following month. After a re-perf it comes in at 1,350. The naive read is a 450-barrel bump versus the prior month. The honest read is 1,350 minus the 820 you'd have gotten anyway — 530 barrels of true incremental production. Then next month the baseline says 760 and the well makes 1,180: another 420 incremental. Stack those monthly deltas and you have the number that pays back the AFE.
This is where production-trend analysis earns its keep. The platform already carries the well's own history and can characterize the pre- and post-job slopes separately, so you can see not just the level change but whether the decline reset. A good re-frac or re-perf often flattens the curve for a while, not just lifts it — and a flatter decline compounds into far more recovered volume than a one-time spike.
Watch for the fade
The uplift that matters is the one that holds. Plenty of interventions — especially chemical treatments and cleanouts — produce a sharp, short-lived response as the wellbore flushes out, then settle back to the original trend within two or three months. Measured over 30 days, those look like wins. Measured over six months, they barely cover the cost.
Set an alert on the well after the job so you're notified when its production changes materially or when it breaks from the new post-workover trend. If the well starts declining faster than the fresh baseline, the treatment is wearing off and you can decide whether to repeat it or move on. This is the same outlier-detection logic used to flag any well departing from its own trend — pointed at the specific question of whether an intervention is durable.
Compare the job across the book
Once you can score one workover cleanly, do it across every well an operator has intervened on. The record lets you line up incremental barrels against job type and vintage: which re-perfs paid back, which chemical programs fizzled, which artificial-lift conversions held. Over a book of dozens of interventions, patterns emerge that no single well shows — maybe re-perfs in one bench consistently return 8,000+ incremental barrels while the same job two counties over returns half that.
You can also benchmark the response against offset wells that got the same treatment. If your uplift is running well below what nearby operators pulled from comparable jobs, that's a design or execution question worth asking before you write the next AFE.
The bottom line
A workover isn't judged by whether production went up — it's judged by how many barrels it added above the decline you'd have gotten for free, and how long those barrels keep coming. Rebuild the baseline, subtract it from actuals, and watch the post-job trend for the fade. Do that against the full production record and you stop paying for bumps and start paying for uplift.