The short version
Almost every acquisition announcement promises that the combined sales forces will sell more together than they did apart, and most never deliver it. I built a program that did, across 31 business units in the NTT Group, and in 22 months it produced more than $200M in contract value and more than $1B in pipeline.
- McKinsey's research found that almost 70 percent of mergers fail to achieve their expected revenue synergies. Cost synergies come in far more reliably, because management can deliver them on its own.
- Cross-selling between sister companies fails for structural reasons: each company is paid on its own P&L, and nobody has agreed who leads a shared deal. Enthusiasm fixes neither.
- At NTT, executives from every operating company agreed in the first 90 days who is prime, secondary and tertiary on a deal, before the field had to argue about it.
- A cross-entity compensation model made a seller whole wherever a deal booked, which turned the rules from requests into something people acted on.
- Deal intake, a review board, a permanent record of the originating company and Salesforce data made the program a mechanism.
1. Revenue synergy is the promise most likely to stay on the slide.
In 2004 the McKinsey Quarterly published a study called "Where mergers go wrong", by Scott Christofferson, Robert McNish and Diane Sias. Its central finding is lopsided: almost 70 percent of the mergers in McKinsey's database failed to achieve the revenue synergies the acquirer expected. Acquirers estimated cost synergies far more accurately, although those were still overestimated by 20 percent or more in about 35 percent of cases.
The people who write the synergy slide deserve some credit, because the logic on it is usually sound. Two companies with different strengths often do share customers who would rather buy more from one group than manage two vendors, and the sellers in each company do have relationships the other could use. The slide is right about the opportunity, and where it goes wrong is in assuming that the opportunity will be taken.
What explains the failure is the gap between the two kinds of synergy: a cost synergy is something management can do by itself, such as merging two finance teams. Revenue synergy needs hundreds of sellers, in companies that keep their own profit and loss statements and their own quotas, to change whom they call and to share accounts and credit with people they have never met. No executive can order that into existence. A seller who brings a sister company into a deal and gets nothing on their own number has been asked to work for free, and a sensible seller declines.
My position, after building one of these programs, is that revenue synergy is a design problem. It gets delivered when the rules of engagement and the pay are designed before the first conflict, and when the program runs as a mechanism that does not depend on anyone's heroics.
2. The opportunity after the acquisition was a sales force already inside our accounts.
When NTT acquired Dimension Data, most people saw a new parent company, while what I saw was hundreds of NTT sellers who already had relationships in accounts we could serve. So I asked to own NTT as an account, and partnered with NTT Communications to sell into their customers.
That work won the first NTT Group Cross-Sell Award, and afterward I was asked to build the cross-sell program for the Americas, reporting into the NTT Global Business Office in Tokyo.
The sequence matters to the argument: I proved it with one sister company before building the program, so the program started from a working example rather than a mandate.
Reporting into Tokyo mattered too, because Dimension Data had been a South African company and NTT is Japanese. Working inside both taught me to operate in cultures that are not my own, including how to spot the real decision-maker when that person says little. In a program that crosses legal entities, knowing who actually decides is most of the work of winning executive support.
3. Start from the portfolio, so every company can see what it gains by selling the others.
The Americas program had to bring 31 business units across Dimension Data, NTT Communications and NTT DATA into one motion. Those were separate legal entities with separate P&Ls, and every one of them had a reason to protect its own accounts.
We started from the portfolio instead of the org chart, so technical and product leaders from every operating company mapped who does what, and we built and published a solution set of natural extensions: each company's go-to-market, extended by the naturally aligned fits from the others.
Starting there did two things, and the first was to give every seller a concrete answer to the question a cross-sell program lives on, which is what the other companies sell that my customer would actually buy from us. Just as important, it took the fight over ownership out of the abstract. Once the overlap is on paper, the conversation shifts from whose account it is to which company leads on which capability, and that second question has an answer.
4. Rules of engagement have to be settled at the top, once, before the first fight.
In the first 90 days, executive leadership from the operating companies aligned on the program, and the matrix of who leads what was set. On every deal, the matrix answered who is prime, who is secondary and who is tertiary, and which company leads on which capability.
Timing is the whole point, because the conflicts in a cross-sell program are predictable: two companies claim the same account, or a sister company wants the lead on a deal where its part is small. If the rule for each conflict is decided after it starts, every decision feels personal and somebody's leader has to escalate it. Reaching consensus one deal at a time does not work. A rule agreed once at the top settles the question before the field has to argue about it.
My lesson in the cost of assuming support came the hard way, on a $42M bid where we closed the price gap to $15,000 and still lost, because I had never secured the partner's executive support. Since then I have won a partner's executive commitment before the bid instead of counting on it. Sister companies inside a group deserve the same treatment as partners outside it, because legally and financially that is what they are.
| Role on the deal | Who holds it | What it owns | How credit works |
|---|---|---|---|
| Prime | The company whose capability is at the center of what the customer is buying | The pursuit, the customer relationship for the deal, and the contract | Books the deal |
| Secondary | The company with the next-largest part of the solution | Its own scope, and a seat in every customer meeting about it | Its seller is made whole under cross-entity pay |
| Tertiary | A company adding a smaller capability | Specialists on call for its part | Credited for its part |
| Originating company | Whoever found the opportunity, whatever role it ends up in | A permanent mark on the opportunity record | Its seller is made whole even when another company is prime |
Consider, hypothetically, how the matrix works on one deal, in which a seller at the networking company in a group like this learns that a regional hospital wants to move its core applications into a managed cloud, which a sister company does better. Without rules, the seller has two poor choices: sell a weaker version from their own catalogue, or hand the lead over and hope someone remembers who found it. With the matrix, I would have the seller log it through deal intake the same day. The review board would qualify it, the matrix would make the cloud company prime and the networking company secondary, and the networking company would stay recorded as the originator for the life of the opportunity.
5. Pay decides whether the rules are followed, so a seller has to be made whole wherever the deal books.
Rules of engagement without pay behind them are requests, because if the prime company books the deal and the originating seller's number shows nothing, that seller has every reason to keep the next opportunity at home, even when a sister company would serve the customer better. Few sellers will say that out loud, but they simply stop bringing deals forward.
To fix that, I designed a cross-entity compensation model that made a seller in one company whole when the deal booked in another. It is the piece I would not compromise on in any integration, and it is often the piece left until later, because changing pay across legal entities needs finance and every operating company's leadership to agree at once. Leaving it until later means every rule written in the meantime is optional.
Recording the originating company was the other half, and we always recorded which company originated an opportunity, even when another company ended up prime on the contract. That record protects the seller's credit, and it shows leadership where cross-sell opportunities actually come from, which is the number you need when deciding where to invest next.
Later, at AWS, I saw the same principle work from the other side of a partnership, when MongoDB changed its reps' pay to annualized consumption credit and I used the change to put both field teams on the same side of the account. A change in how people were paid changed how they sold.
6. The program ran on a cadence, so no deal depended on someone forcing it through.
A deal intake process qualified every cross-sell opportunity and assigned a cross-company team to pursue it. An opportunity review board qualified each one before it counted, so the pipeline we reported was a pipeline we could defend. Both matter more in a cross-sell program than in a single sales team, because a number built from 31 business units is easy to inflate and hard to check.
In 2016 and 2017 we ran a program called Data Driven Decisions, tied into Salesforce reporting, to allocate cross-sell resources where they would produce. With 31 business units each wanting support, data is what kept allocation from turning into a contest of who argued hardest.
Enablement ran on a fixed rhythm: regional quarterly business reviews brought in leaders from every operating company and business unit. Twice a month we ran lunch-and-learns in sales offices across the country, pulling in the regional NTT teams and not just the host company, and we gave each company's sales and technical teams marketing packs they could use. I also wrote the One NTT messaging the group adopted, so customers heard one account of the group whichever company's seller was in the room.
None of these pieces is clever on its own, but together they meant an opportunity moved from discovery to a cross-company team without anyone having to push it through by force of personality, which is the only way a program across 31 business units can keep running.
7. Amazon has a name for what made it work: a mechanism.
At the time I did not have a word for this, but Amazon has one, from its saying that good intentions don't work, mechanisms do. Gregor Hohpe, writing on the AWS Enterprise Strategy blog, describes mechanisms as complete processes made of a tool, adoption of that tool, and inspection to check that the mechanism works.
| Part of a mechanism | What Amazon means by it | What it was in the NTT program |
|---|---|---|
| Tool | The process or system that does the work | The solution set, the prime, secondary and tertiary matrix, the cross-entity compensation model, and deal intake |
| Adoption | Getting people to actually use it | Executive alignment in the first 90 days, twice-a-month lunch-and-learns, marketing packs, and the One NTT messaging |
| Inspection | Checking the results and adjusting | The opportunity review board, Data Driven Decisions on Salesforce, and regional quarterly reviews |
Read against that definition, a cross-sell program launched with a kickoff and an award is a good intention. It works while the executive who launched it keeps pushing, and it fades when that executive's attention moves to the next priority. The NTT program had all three parts, and that is why I credit the results to the design rather than to any single deal or person, including me.
My region was named Top Region in the NTT Group in 2016, 2017 and 2018. Of the program's own results, the pipeline figure is the one I would point a new integration leader to. Contract value tells you the program closed deals. A pipeline that size tells you that sellers across 31 business units kept bringing opportunities forward, and sellers only do that when the rules and the pay make it safe.
8. What this means for leaders integrating sales forces
- Write the rules of engagement in the first 90 days. Get executive leadership from every operating company to agree who is prime, secondary and tertiary, and who leads on which capability, before the field has its first argument.
- Map the portfolio before you redraw the org chart. Have technical and product leaders publish what each company sells that naturally extends the others. Sellers cross-sell what they can explain to a customer.
- Fix the pay before you ask for the behavior. Make a seller whole wherever the deal books. Until that is true, every rule you write is a request.
- Record the originating company on every opportunity, permanently. It protects the seller's credit, and over time it tells you which relationships are producing.
- Qualify every cross-sell opportunity once, the same way. Deal intake and a review board keep a pipeline built from many companies honest, so a forecast means the same thing in every one of them.
- Inspect with data and move resources to what produces. Quarterly reviews with every company in the room, and enablement on a fixed rhythm, keep the mechanism in use after the announcement is forgotten.
Revenue synergy will keep appearing on acquisition slides, and it should, because the opportunity is usually real. Whether it ever reaches the numbers depends on someone designing the rules and the pay before the first conflict, and then running the program as a mechanism that does the work without heroics.
Sources
- Scott A. Christofferson, Robert S. McNish and Diane L. Sias, "Where mergers go wrong", McKinsey Quarterly, 2004 Number 2: mckinsey.com. The figures (almost 70 percent missing expected revenue synergies, cost synergies overestimated by 20 percent or more in about 35 percent of cases) are also reported in "New McKinsey research challenges conventional M&A wisdom", Strategy & Leadership (Emerald), 2004 and CFO.com, "Fool's Gold".
- Amazon's saying "Good intentions don't work, mechanisms do" and its definition of a mechanism (tool, adoption, inspection): Gregor Hohpe, "Strategy Is a Winding Road. Mechanisms Keep You on Track", AWS Enterprise Strategy blog, 15 July 2021.
- NTT Americas cross-sell program results, the NTT Group Cross-Sell Award and Top Region 2016 to 2018: NTT records, by reference.