A vendor's AI ROI study says 244%. The composite organisation it models was never your rollout.

On 7 October 2026, Outreach published a Forrester-commissioned Total Economic Impact study claiming a 244% three-year return on investment, a payback period under six months and $19.2m in benefits, present value, for a modelled "composite organisation" built from its own reference customers. The number is real, Forrester's name is attached to it, and the methodology behind it is standard across the sales technology industry.

None of that makes the number yours. A composite organisation is built to describe an average of a handful of already-successful accounts, not the rollout your team is about to run, and the gap between those two things is the part no commissioned study is built to tell you.

What did Outreach's new study actually say?

Forrester modelled a hypothetical composite organisation from Outreach's reference customers and reported 244% three-year ROI, under six months' payback, $19.2m in benefits and a 20% year-one productivity lift on admin work.

Outreach's own newsroom page, dated 7 October 2026, lists the headline figures plainly: 244% return on investment over three years, payback in under six months, $19.2m in benefits present value, $13.6m net present value, and a 20% productivity lift on administrative sales activities in year one. These are Outreach's figures, from a study it commissioned, describing the composite organisation Forrester built to represent the customers it interviewed.

What is a "composite organisation," and why does almost every vendor publish one?

Forrester's Total Economic Impact method interviews a small number of a vendor's own customers and models their outcomes onto one hypothetical organisation, which is how it turns a handful of interviews into one headline multiple.

Forrester's own description of the Total Economic Impact methodology is explicit about the mechanics: it builds financial models based on customer interviews, then represents the findings as one composite organisation standing in for the group. This is not something specific to Outreach. The same structure sits behind the commissioned TEI studies Salesforce, HubSpot and most of the revenue technology category have published over the years. It is an industry format, not a vendor's trick, and that is exactly why it is worth understanding rather than just quoting.

Why doesn't a composite organisation's number transfer to your own rollout?

The interviewed customers are the vendor's own reference accounts, already succeeding, so the model is built from a sample that excludes a struggling rollout before the study even starts.

Three mechanical reasons, not an accusation against Outreach or Forrester, since this is how the format works for every vendor that uses it.

First, selection. The customers interviewed for a commissioned study are the vendor's own reference accounts, the ones willing to get on a call and talk about a deployment that went well. A team still fighting adoption six months in is not in that sample, so the composite organisation is averaged from outcomes that have already cleared the bar a new buyer has not cleared yet.

Second, averaging. One hypothetical organisation collapses different account sizes, team structures and starting process maturity into a single payback window and a single productivity percentage. The organisation the number describes does not exist anywhere. It is closer to a tailor's dummy built to an averaged measurement than to the body of any one customer trying the suit on.

Third, attribution. A year-one productivity lift bundles the software with whatever training, process redesign and management attention the customer's own team put in alongside it. The study reports the combined result as the tool's number, because that is what the format measures, not because anyone is hiding the rest.

What should you actually do with a number like 244%?

Use the study to understand where a vendor claims the value comes from, then test that mechanism against your own baseline data instead of citing the multiple in a business case.

Our view, from building the business case for automation spend with clients rather than reading one off a vendor's newsroom page: a commissioned TEI study is a reasonable map of where a tool's value is supposed to come from, admin time saved, pipeline visibility, faster sequencing, and a poor source for how much of it you will actually get. In our experience, the only number worth carrying into a board deck is one measured from your own CRM data over a defined window, before and after rollout, using the same weekly update-rate discipline we apply to any pipeline number we are asked to defend.

To be fair to Outreach: sequencing and conversation intelligence tools of this kind genuinely do save reps time on administrative work, and the direction of the claim, less manual admin and more selling time, is plausible on its face. The study's fault is not dishonesty. It is that a composite organisation was never going to be a stand-in for the one organisation actually reading the report.

We wrote the same caution about vendor claims the other way round two weeks ago, when Microsoft's Sales Development Agent entered a free preview window too short to judge its qualification calls against real outcomes, in the real test takes longer than that. A generous free trial and a generous commissioned study fail for the same reason: the number on offer was never measured against your deployment, because it could not have been.

A 244% ROI figure is not a reason to buy, and it is not a reason to disbelieve either. It is a map of where a vendor expects value to show up. The only number that belongs in your own board deck is the one you measure from your own data, the same way we build the pipeline coverage check and the forecast you can defend from a client's own history rather than a published benchmark. If you are building the business case for an automation purchase and want help separating the vendor's mechanism from your own measured baseline, see our AI automation work.